LTE Mobility and Throughput - KPI Analysis & Optimization Workshop STUDENT BOOK LZT1381950 R1A LZT1381950 R1A LTE Mobility and Throughput - KPI Analysis & Optimization Workshop DISCLAIMER This book is a training document and contains simplifications. Therefore, it must not be considered as a specification of the system. The contents of this document are subject to revision without notice due to ongoing progress in methodology, design and manufacturing. Ericsson shall have no liability for any error or damage of any kind resulting from the use of this document. This document is not intended to replace the technical documentation that was shipped with your system. Always refer to that technical documentation during operation and maintenance. © Ericsson AB 2017 This document was produced by Ericsson. • The book is to be used for training purposes only and it is strictly prohibited to copy, reproduce, disclose or distribute it in any manner without the express written consent from Ericsson. This Student Book, LZT1381950, R1A supports course number LZU1082504. -2 - © Ericsson AB 2017 LZT1381950 R1A Table of Contents Table of Contents 1 LTE MOBILITY PERFORMANCE AND RELATED PARAMETERS ............................................................................ 11 1 INTRODUCTION ............................................................................ 12 2 IDLE MODE MOBILITY MANAGEMENT ........................................ 12 2.1 PRIORITY BASED CELL RESELECTION ................................... 14 2.2 SPEED-DEPENDENT SCALING OF CELL RESELECTION ........ 16 2.3 PARAMETER RELATED TO IDLE MODE MOBILITY .................. 17 2.3.1 CASE 1: PRIORITY BASED CELL RESELECTION EXAMPLE ............................................................................................ 18 2.3.2 CASE 2: 3CC CA EXAMPLE..................................................... 18 2.3.3 CASE 3: INITIAL TRAFFIC BALANCING .................................. 20 2.3.4 CASE 4: RESELECTION PARAMETER EXAMPLE .................. 20 2.4 STICKY CARRIER FOR IDLE MODE MOBILITY DURING IFLB 23 3 EUTRA CONNECTED MODE MOBILITY KPI................................. 25 3.1 LTE EVENT REVIEW................................................................... 25 3.1.1 MOBILITY WITH “MOBILITY CONTROL AT POOR COVERAGE” FEATURE ...................................................................... 27 4 MOBILITY RELATED ISSUES ANALYSIS ...................................... 28 4.1 HANDOVER FAILURE ISSUES ................................................... 29 4.2 HANDOVER PREP FAILURE- NEIGHBOR CELL LOAD ISSUES ................................................................................................ 30 4.3 HANDOVER FAILURE, TARGET CELL LICENSE AND ADMISSION ISSUE .............................................................................. 31 4.4 HANDOVER PREP FAILURE, OTHER CHECKS ........................ 32 5 HANDOVER EXECUTION FAILURE ISSUES AND COUNTERS ... 35 5.1 INTRA/INTER HANDOVER EXECUTION FAILURE, POSSIBLE CAUSE .............................................................................. 35 LZT1381950 R1A © Ericsson AB 2017 -3 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop 5.2 HANDOVER EXEC FAILURE, POOR DL SINR ISSUES RELATED PARAMETER ...................................................................... 36 5.3 HANDOVER EXEC FAILURE, UL RSSI ISSUES RELATED PARAMETER ....................................................................................... 36 5.4 HANDOVER EXEC FAILURE, TARGET CELL RACH ISSUE...... 37 5.5 OVERSHOOTING CELL .............................................................. 39 5.6 HANDOVER PREP/EXEC FAILURE, OTHER IMPORTANT CHECK ................................................................................................. 39 6 IRAT HANDOVER HO OPTIMIZATION .......................................... 40 7 MCPC RELATED COUNTERS AND PARAMETERS ..................... 41 8 INTER FREQUENCY LOAD BALANCING COUNTERS AND PARAMETERS ..................................................................................... 42 9 LTE FEATURE SUPPORTING MOBILITY KPI: .............................. 44 10 AUTOMATED MOBILITY OPTIMIZATION .................................... 45 10.1 PARAMETER RELATED TO AMO FEATURE ........................... 46 10.1.1 COUNTERS, EVENTS AND PARAMETERS AMO ................. 48 10.2 CASE STUDY 1, AMO TRIAL TEST .......................................... 49 11 MOBILITY CONTROL AT POOR COVERAGE ............................. 52 11.1 MCPC ASSOCIATED KPI AND COUNTERS ............................. 54 11.2 PARAMETERS ASSOCIATED TO MCPC FEATURE ................ 54 11.3 CASE 2, POST MCPC IMPACT ON KPIS.................................. 57 12 MULTI-LAYER SERVICE-TRIGGERED MOBILITY ...................... 59 13 SUMMARY.................................................................................... 61 2 DIFFERENT HOSR ISSUES AND IMPROVEMENT - CASE STUDIES ..................................................................................... 63 1 COMMON FAILURE SCENARIO .................................................... 64 1.1 X2 HO PREPARATION PHASE ................................................... 64 1.2 X2 HANDOVER EXECUTION PHASE ......................................... 65 1.3 HO EXECUTION FAILURE EXAMPLE ........................................ 66 -4 - © Ericsson AB 2017 LZT1381950 R1A Table of Contents 1.4 EVENT IDENTIFICATION ............................................................ 67 2 CASE STUDIES .............................................................................. 68 2.1 CASE 1- IRAT HO, OPTIMUM PARAMETER SETTING TO IMPROVE LTE DROP RATE ................................................................ 68 2.2 CASE 2: INTRA FREQUENCY OSCILLATING HANDOVER ....... 70 2.2.1 NETWORK ISSUE .................................................................... 70 2.2.2 CASE 2: INTRA FREQUENCY HO EVENT REVIEW ................ 71 2.2.3 PARAMETERS TUNED ............................................................ 71 2.2.4 IMPACT: KPI IMPROVEMENT ................................................. 72 2.3 CASE 3: HOSR DEGRADED WITH AMO ACTIVATION .............. 73 2.3.1 ISSUE STATEMENT ................................................................. 73 2.3.2 CIO STATISTICS OBSERVATION............................................ 73 2.3.3 CASE 3: CIO = -2 & -3 DISTRIBUTION IN DIFFERENT CLUSTER............................................................................................. 75 2.3.4 CIO = -3 FOR FIRST TIER NEIGHBORS.................................. 75 2.3.5 POTENTIAL REASON OF UNREASONABLE CIO VALUE....... 75 2.3.6 CASE 3: TRAIL SUGGESTION ................................................. 76 2.4 CASE 4: HOSR EXECUTION IMPROVEMENT ON DISTANCE SITE .................................................................................. 77 2.4.1 CASE 4: HOSR EXECUTION IMPROVEMENT WITH CELL RANGE CHANGE ................................................................................ 77 2.5 CASE 5: POOR IRAT SUCCESS RATE TOWARDS POOR WCDMA CELL FOR QCI8 .................................................................... 78 2.5.1 CASE 5: POOR IRAT PARAMETERS CHECK ......................... 79 2.5.2 CASE 5: POOR IRAT SUCCESS RATE TOWARD POOR WCDMA CELL FOR QCI8 .................................................................... 79 2.6 POOR IRAT HOSR – OTHER REASONS & SOLUTIONS ........... 80 2.6.1 IRAT IMPROVEMENT COUNTER TO BE CHECKED .............. 80 2.6.2 CASE 6: IRAT IMPROVEMENT WITH ‘A1A2SEARCHTRESHOLDRSRP’ ....................................................... 82 LZT1381950 R1A © Ericsson AB 2017 -5 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop 2.6.3 HIGH SAMPLES FOR ‘PMCRITICALBORDEREVALREPORT’ ............................................... 82 2.7 CASE 7: KPIS (HOSR, RET., THP.) IMPROVEMENT WITH IFHO PARAMETER SETTING ............................................................. 83 2.8 CASE 8: COVERAGE ISSUE, CRSGAIN TUNING ...................... 86 2.9 CASE 9: HANDOVER FAILED IN PREPARATION PHASE ......... 87 2.10 CASE 10: PCI CONFUSION ...................................................... 89 2.11 CASE 11: ~100% HO EXECUTION FAILURES ......................... 90 2.11.1 CASE 11- RRC HO EXECUTION FAILURES DUE TO PCI COLLISION .......................................................................................... 91 3 ANR: MOM – ANR CREATED ........................................................ 91 3.1 HANDOVER EVENT TRIGGERED .............................................. 92 3.2 PERIODICAL ANR (INTER-FREQUENCY AND IRAT HO) .......... 93 3.3 ANR PARAMETER EUTRAN ....................................................... 95 3.4 OBSERVABILITY ......................................................................... 98 3.5 ANR RELATED COUNTERS ....................................................... 98 3.6 SELF-ORGANIZED NETWORK & ANR FEATURES ................... 99 3.7 AUTOMATED NEIGHBOUR RELATION (PCI CONFLICT IMPACT)............................................................................................... 99 3.8 AUTOMATED NEIGHBOUR RELATION (PCI CONFLICT HANDLING)........................................................................................ 100 3.9 AUTOMATED NEIGHBOUR RELATION (PCI CONFLICT DETECTION DRX) ............................................................................. 101 3.10 AUTOMATED MOBILITY OPTIMIZATION (COORDINATION WITH ANR) ........................................................................................ 102 4 CS FALLBACK.............................................................................. 103 4.1 CSFB RELATED FEATURES .................................................... 104 4.2 CSFB CASES FOR LONG CALL SETUP TIME ......................... 105 4.3 CASE 1: CSFB CALL SETUP TIME DUE TO CORE SIGNALING ........................................................................................ 105 4.4 CSFB FAILURE DUE TO LONG CELL RESELECTION TIME ... 106 -6 - © Ericsson AB 2017 LZT1381950 R1A Table of Contents 4.5 SIB READING DURING CELL RESELECTION ......................... 106 4.6 FAILURE DUE TO MULTIPLE RRC CONNECTION REQUEST .......................................................................................... 107 4.7 FAILURE DUE TO MULTIPLE RRC CONNECTION REQUEST .......................................................................................... 107 4.8 FAILURE DUE TO BAD UTRAN COVERAGE ........................... 108 4.9 CASE 2: UE FAIL TO RETURN TO LTE AFTER CSFB TO UTRAN ............................................................................................... 109 4.9.1 CASE 2: UE FAIL TO RETURN TO LTE AFTER CSFB TO UTRAN ............................................................................................... 109 4.9.2 CASE 2: UE FAIL TO RETURN TO LTE AFTER CSFB TO UTRAN ............................................................................................... 110 4.10 CASE 3: UE TOOK MORE TIME WHEN RETURN BACK TO LTE NW1 ............................................................................................ 111 5 SUMMARY ................................................................................... 113 3 LTE INTEGRITY KPIS PERFORMANCE AND RELATED PARAMETERS .......................................................................... 115 1 INTRODUCTION .......................................................................... 116 1.1 AVERAGE UE DOWNLINK THROUGHPUT .............................. 116 1.1.1 DOWNLINK DRB TRAFFIC VOLUME..................................... 117 1.1.2 AVERAGE UE UPLINK THROUGHPUT ................................. 119 1.1.3 UPLINK DRB TRAFFIC VOLUME ........................................... 120 1.1.4 EUTRAN LATENCY KPIS ....................................................... 121 1.1.5 EUTRAN PACKET LOSS KPIS ............................................... 123 2 THROUGHPUT OPTIMIZATION STEPS ...................................... 127 2.1 DL THROUGHPUT ISSUES INVESTIGATION .......................... 127 2.2 UPLINK THROUGHPUT INVESTIGATION ................................ 128 2.3 REASONS FOR POOR DL THROUGHPUT ................................ 129 2.4 REASONS FOR POOR UL THROUGHPUT ................................ 130 2.5 THROUGHPUT TESTING INVESTIGATION ............................. 131 LZT1381950 R1A © Ericsson AB 2017 -7 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop 3 LOW THROUGHPUT- THE COUNTERS TO BE CHECKED ........ 132 3.1 PRB UTILIZATION ..................................................................... 132 3.2 WAYS TO REDUCE PRB UTILIZATION .................................... 133 3.3 ESSENTIAL PARAMETERS ...................................................... 134 3.4 OTHER COUNTERS FOR THROUGHPUT ANALYSIS ............. 134 3.5 UL INTERFERENCE .................................................................. 135 3.6 UL RLC NACK ........................................................................... 136 3.7 POWER RESTRICTED TRANSPORT BLOCK IN UL ................ 136 3.8 CQI RANGE ............................................................................... 137 3.9 MIMO RANK DISTRIBUTION USAGE ....................................... 138 3.10 MODULATION SCHEME USAGE DL ...................................... 138 4 THROUGHPUT OPTIMIZATION- UE ESSENTIAL CHECKS ....... 139 4.1 QOS FRAMEWORK .................................................................. 140 4.2 SCHEDULING ALGORITHM SCHEDULER CONFIGURATION ............................................................................. 142 5 FEATURE AFFECTING UL/DL THROUGHPUT ........................... 144 5.1 256 QAM DL .............................................................................. 145 5.2 64 QAM UPLINK ........................................................................ 148 5.3 64 QAM UL PARAMETERS ....................................................... 150 5.4 ERICSSON LEAN CARRIER ..................................................... 151 5.5 CARRIER AGGREGATION ....................................................... 153 5.6 POTENTIAL THROUGHPUT ..................................................... 154 5.7 UPLINK CARRIER AGGREGATION .......................................... 156 5.8 UPLINK CARRIER AGGREGATION CONFIGURATION ........... 156 5.9 ADVANCED CARRIER AGGREGATION (FAJ 801 0568).......... 158 5.9.1 INTER-ENB CARRIER AGGREGATION PERFORMANCE .... 160 5.9.2 INTER-ENB CARRIER AGGREGATION PARAMETERS ....... 161 -8 - © Ericsson AB 2017 LZT1381950 R1A Table of Contents 5.10 MULTI-CARRIER LOAD MANAGEMENT (FAJ 801 0424) ....... 162 5.11 PRESCHEDULING .................................................................. 162 5.11.1 PRESCHEDULING PARAMETERS ...................................... 164 5.12 LTE TRANSMISSION MODES DOWNLINK TRANSMISSION MODES .............................................................................................. 165 5.12.1 WHICH MIMO MODE IS BEST? ........................................... 167 6 VOLTE IMPACT ON BB THROUGHPUT & DBS .......................... 168 7 SUMMARY ................................................................................... 170 4 ISSUE ANALYSIS, IMPROVEMENTS AND CASE-STUDIES FOR INTEGRITY KPIS .............................................................. 171 1 CASE-1: DL THROUGHPUT IMPROVEMENT ............................. 172 2 CASE-2 UL-THROUGHPUT IMPROVEMENT .............................. 173 3 CASE-3: LTE LATENCY ISSUE.................................................... 174 3.1 CASE-3: INVESTIGATION ANALYSIS ...................................... 175 3.2 CASE-3 INVESTIGATION FOR SERVICE SPECIFIC DRX ....... 176 3.3 CASE-3 DRX PROFILE PARAMETER OPTIMIZATION ............ 176 4 CASE-STUDY 4: LOW THROUGHPUT DURING HIGH LOAD ..... 178 4.1 EXERCISE 1 – ANALYZE ROP FILES ...................................... 179 4.1.1 EXERCISE 1 – ANSWER........................................................ 180 4.2 EXERCISE 2: POSSIBLE CAUSE FOR THROUGHPUT DEGRADATION ................................................................................. 180 4.2.1 EXERCISE 2 – ANSWER UL/DL POOR RF COUNTERS ....... 181 4.2.2 UL NOISE AND INTERFERENCE .......................................... 182 4.2.3 PDCP LAYER COUNTERS..................................................... 183 4.2.4 RLC LAYER COUNTERS ....................................................... 183 4.2.5 MAC LAYER COUNTERS....................................................... 184 4.2.6 PDCCH CONGESTION COUNTER ........................................ 185 4.2.7 UPLINK-DOWNLINK PRB UTILIZATION ................................ 185 LZT1381950 R1A © Ericsson AB 2017 -9 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop 4.2.8 COUNTERS FOR UES SCHEDULED PER TTI: ..................... 186 4.3 EXERCISE 3: POSSIBLE ROOT CAUSE .................................. 187 4.3.1 COUNTER PMRADIORECINTERFERENCEPWR ANALYSIS-RESULTS ........................................................................ 187 4.3.2 COUNTER PMRADIOUEREPCQIDISTR ANALYSISRESULTS ........................................................................................... 188 4.3.3 COUNTER PMRADIOUEREPCQIDISTR2 ANALYSISRESULTS ........................................................................................... 188 4.3.4 PDCP LAYER COUNTER ANALYSIS-RESULTS ................... 189 4.3.5 RLC LAYER COUNTER ANALYSIS-RESULTS ...................... 190 4.3.6 COUNTER ‘PMMACHARQ’ ANALYSIS-RESULTS FOR UL/DL 190 4.3.7 MAC HARQ COUNTERS ANALYSIS-RESULT....................... 191 4.3.8 COUNTER PMPDCCHCCEUTIL ANALYSIS-RESULTS ......... 191 4.3.9 SE UTILIZATION ANALYSIS-RESULTS ................................. 192 4.4 EXERCISE SUMMARY .............................................................. 192 4.5 WAY FORWARD ....................................................................... 193 5 SUMMARY OF CHAPTER 4 ......................................................... 194 5 ABBREVIATIONS .......................................................................... 195 6 INDEX ............................................................................................ 207 7 TABLE OF FIGURES ..................................................................... 215 - 10 - © Ericsson AB 2017 LZT1381950 R1A LTE Mobility performance and related parameters 1 LTE Mobility performance and related parameters Objectives After completion of this chapter the participants will be able to: 1 Analyze LTE Mobility performance and related parameters. 1.1 Explain Parameter related to Idle mode mobility. 1.2 Describe different LTE mobility KPIs & counters for X2HO, S1HO, IFHO, IRATHO. 1.3 Validate Mobility related parameter affecting different KPIs. 1.4 Evaluate the steps for optimization of these KPI 1.5 Analyze features with related parameters which improve LTE Mobility KPIs Figure 1-1: Objective of Chapter 1 LZT1381950 R1A © Ericsson AB 2017 - 11 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop Introduction 1 This chapter introduces the LTE mobility KPIs (Key Performance Indicators) with their associated counters, and parameters that might affect them directly. It will also review some of the probable reasons for failure and the subsequent steps that may be undertaken to optimize the network. The Optimization of all the KPIs (for Accessibility, Retainability, Integrity, Mobility) begins with the collection of network statistics. These performance statistics are put in formulae to create meaningful Key Performance Indicators (KPIs) and evaluated against guiding thresholds. Sometimes Performance Recordings (i.e. Cell Trace and UE Trace) are also initiated and evaluated. In some cases, even drive tests are performed. The data is then analyzed. Based on this analysis the optimizer is expected to recommend on how the performance can be further improved. After the changes have been implemented, they are again verified with network performance data (statistics and recordings). The whole process is repeated until the desired KPI target values are reached. In this chapter, focus will be on Idle mode and Connected mode mobility. Idle Mode Mobility Management 2 Idle mode mobility optimization is done to optimize cell coverage, and to correctly define frequency layering to prioritize one cell/frequency over the other. Idle mode optimization is based on the cell selection and cell reselection parameters. These parameters are defined as the attributes in ‘EUtranFreqRelation’, ‘UtranFreqRelation’, ‘EUtranCellRelation’, ‘EUtranCellFDD’ and ‘EUtranCellTDD’ Managed Objects (MOs). Although the Idle Mode Mobility does not affect the pre-defined KPIs, it is still an important area to consider for the best performance out of a network.Figure 12 explains cell selection process and Figure 1-3 illustrates cell reselection process. Following are the important parameters, to be optimized for cell selection and reselection. ‘pMax’ = 1000 { -30.33, 1000} Calculates the parameter ‘Pcompensation’ (defined in 3GPP TS 36.304), at cell reselection to an E-UTRAN Cell. Value 1000 means the parameter is not sent in the system information block. Unit dBm. ‘qOffsetFreq’ = 0 { -24, -22, -20, -18, -16, -14, -12, -10, -8, -6, -5, -4, -3, -2, -1, 0, 1, 2, 3, 4, 5, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24} Frequency specific offset for E-UTRAN frequencies used in connected mode. In idle mode, the negative value of this offset is used. Specification: 3GPP TS 36.331, Unit: 1 dB. ‘qQualMin’ = 0 { -34.-3, 0} Parameter ‘Qqualmin’ in TS 36.304. Value 0 means that it is not sent and UE applies in such case the (default) value of negative infinity for ‘Qqualmin’. Sent in SIB3 or SIB5. - 12 - © Ericsson AB 2017 LZT1381950 R1A LTE Mobility performance and related parameters ‘qRxLevMin’ = -140 { -140, -44} The required minimum received Reference Symbol Received Power (RSRP) level in the (E-UTRA) frequency for cell reselection. Corresponds to parameter ‘Qrxlevmin’ in 3GPP TS 36.304. This attribute is broadcast in SIB3 or SIB5, depending on whether the related frequency is intra- frequency or inter-frequency. Unit: 1 dBm. S-criterion measured RSRP Srxlev = Qrxlevmeas – qrxLevMin + qrxLevMinOffset – Pcompensation* > 0 Squal = Qqualmeas – (qqualMin + qqualMinOffset) Parameters measured RSRQ Srxlev > 0 AND Squal > 0 *Pcompensation = max(pmaxServingCell – P;0) Parameter UE Power: 23 dBm S>0 S>0 S>0 S>0 S>0 S>0 S>0 S>0 S>0 S>0 S>0 S>0 S>0 S>0 S>0 Figure 1-2: Review: Cell Selection (S-Criterion) 0 to 24 dB R(serving) = Qmeas(s) + qHyst R(neighbor) = Qmeas(n) - Qoffset R(neighbor) Rank 2 R(neighbor) Rank 1 R(neighbor) Rank 3 Qoffset qOffsetCellEUtran: Cell individual offset in the intra-frequency and equal priority inter-frequency cell ranking criteria qOffsetFreq: Frequency specific offset in the equal priority interfrequency cell ranking criteria ‘tReselectionEutra’ Qmeas: RSRP measurement quantity used in cell reselections Figure 1-3: Cell Reselection (R-Criteria) LZT1381950 R1A © Ericsson AB 2017 - 13 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop ‘qHyst(s)’ is the hysteresis value ‘qHyst’ that is read in the system information of the serving cell. It prevents too frequent reselection back and forth between cells of nearly equal rank. When a neighboring cell is ranked as better than the serving cell (that is, Rn > Rs) during a time interval ‘tReselectionEutra’, the UE performs a cell reselection to the better ranked cell. ‘Qoffset’ is an offset in the cell ranking criterion of neighbor E-UTRA cells. It consists of a cell individual part and a frequency specific part. The frequency specific part applies to equal priority inter-frequency cells only. According to 3GPP: • For intra-frequency: Equals to ‘Qoffset’, if ‘Qoffset’ is valid, otherwise this equals to zero. • For inter-frequency: Equals to ‘Qoffsets,n’ plus ‘Qoffsetfrequency’, if ‘Qoffsets,n’ is valid, otherwise this equals to ‘Qoffsetfrequency’. • In Ericsson’s network: ‘Qoffsets,n’ = ‘qOffsetCellEUtran’ and ‘Qoffsetfrequency’ = ‘qOffsetFreq’. By tuning ‘qHyst’, ‘qoffset’ and ‘tReselectionEutra’ parameter an operator can optimize the cell reselection procedure as show in Figure 1-4 below. RSRP sIntraSearch sNonIntraSearch qHyst(s) Qmeas(n) R(n) qoffset(s) R(s) Qmeas(s) tReselectionEutra time Cell reselection Figure 1-4: Cell Reselection Evaluation Process 2.1 Priority based Cell Reselection When multiple E-UTRA and/or inter-RAT frequencies are used in a network, priority based cell reselection should be applied. A Network can use ‘cellreselectionPriority’ parameter to prioritize one frequency over the other and optimize idle mode traffic as per the layering plan. The priority based reselection can be used for IRAT frequency as well, but the equal priority for IRAT should be avoided. - 14 - © Ericsson AB 2017 LZT1381950 R1A LTE Mobility performance and related parameters The UE only performs cell reselection evaluation for the E-UTRA and inter-RAT frequencies that are given in system information and for which a cell reselection priority is provided. If the UE finds an inter-frequency carrier or an inter-RAT frequency with a cell reselection priority higher than the frequency where the UE is camping, the UE attempts to reselects a cell on that frequency. Cell reselection occurs if the UE finds a cell with a ‘Srxlev’ value greater than the ‘threshXHigh’ value, or an ‘Squal’ value greater than ‘threshXHighQ’ value for that frequency. The criterion will be used is depends on whether the parameter ‘threshServingLowQ’ is included in SIB3. If the UE finds an inter-frequency carrier with equal priority to the frequency where the UE is camping, the UE performs cell reselection much in the same way as intra-frequency cell reselection. Cell reselection occurs if the UE finds a cell with an ‘Srxlev’ value greater than the ‘threshXHigh’ value, or an ‘Squal value greater than the ‘threshXHighQ’ value for that frequency. Which criterion is used depends on whether the parameter ‘threshServingLowQ’ is included in SIB3. If the ‘Srxlev’ value of the serving cell falls below the ‘threshServingLow’ value, inter frequency carrier or an inter-RAT frequency with cell reselection priority lower than the frequency where the UE is camping. Cell reselection occurs if the UE finds a cell with ‘Squal’ value of the serving cell falls below the ‘threshServingLowQ’ value, the UE attempts to reselect a cell on an inter-freq ‘rxlev’ value greater than the ‘threshXLow’ value or an ‘Squal’ value greater than the ‘threshXLowQ’ value for that frequency. The process is explained in Figure 1-5 below. Neighbor cellReselectionPriority higher than for used freq and = Qmeas(s) + Qoffmbms* Srxlev(n)R(serving) > threshXHigh or Squal(n)+ > threshXHighQ trigger cell reselection to higher prio frequency (E-UTRAN or IRAT) and Frequency 2 Prio 2 Squal(s) < threshServingLowQ Srxlev(s) < threshServingLow Srxlev(n) > threshXLow Frequency 1 Prio 1 and or Squal(n) > threshXLowQ trigger cell reselection to lower prio frequency (E-UTRAN or IRAT) › For RSRQ based reselection threshServingLowQ is included in SIB3 & qQualMin in SIB1 › Configuration of equal priority IRAT frequencies should be avoided! Frequency 3 Prio 3 Figure 1-5: Priority Based Cell Reselection LZT1381950 R1A © Ericsson AB 2017 - 15 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop When the parameter ‘threshServingLowQ’ is included in SIB3, it is required that the parameter ‘qQualMin’, the required minimum RSRQ level (dB) in the cell, is included in SIB1. Important Parameters that correlate to priority based reselection are as following: ‘threshXHigh’ = 4 {0, 62} The threshold used by the UE when reselecting towards the higher priority frequency X from the current serving frequency. Each frequency of E-UTRAN can have a specific threshold. Refer to the parameter ‘thresXHigh’ in 3GPP TS 36.304. Unit: 1 dB. ‘threshXHighQ’ = 2 {0, 31} Parameter ‘ThreshXHighQ’ is defined in TS 36.304 for the ‘EUtranFreqRelation’ that points to the intra frequency EUtranFrequency MO class, Unit: 1 dB. ‘threshServingLowQ’ is included in systemInformationBlockType3. When the parameter ‘threshServingLowQ’ is included in SIB3, it is required that the parameter ‘qQualMin’, the required minimum RSRQ level (dB) in the cell, is included in SIB1. Takes effect: Immediately ‘threshXLow’ = 0 {0...62} The threshold used in reselection towards frequency X priority from a higher priority frequency. Each frequency of E-UTRAN can have a specific threshold. Parameter ‘thresXlow’ is defined in 3GPP TS 36.304. Unit: 1 dB ‘threshXLowQ’ = 0 {0...31} Parameter ‘ThreshXLowQ’ defined in TS 36.304, for the ‘EUtranFreqRelation’ that points to the intra frequency ‘EUtranFrequency’ MO, the attribute is invalid, this attribute does not affect SIB3. Unit: 1 dB. 2.2 Speed-Dependent Scaling of Cell Reselection The usual ‘tReselectionEutra’ and ‘qHyst’ parameters are used in the normal mobility state for the evaluation of the cell reselection criteria. In the medium and high mobility states, the UE applies a scaling factor, decreasing the value of ‘tReselectionEutra’ parameter (‘tReselectionEutraSfMedium’ and ‘tReselectionEutraSfHigh’). In that way, the evaluation period of cell reselection criteria is reduced. In addition, a negative offset is added to the ‘qHyst’ hysteresis value (‘qHystSfMedium’ and ‘qHystSfHigh’) in the cell ranking criteria. It lowers the threshold for the reselection of intra-frequency cells. The criteria for the UE to enter the medium and high mobility states is based on the number of recent cell reselections performed by the UE. A sliding time window is used. The parameter ‘tEvaluation’ determines the duration of the sliding time window. The parameters ‘nCellChangeMedium’ (medium mobility) and ‘nCellChangeHigh’ (high mobility) determine the number of cell reselections the UE performs within the sliding time window. The UE applies an additional timeperiod before reentering the normal mobility state. - 16 - © Ericsson AB 2017 LZT1381950 R1A LTE Mobility performance and related parameters The speed dependent cell reselection is illustrated in Figure 1-6. › Three mobility states: Normal, Medium and High › Based on the no of cell reselections made by the UE – nCellChangeMedium – nCellChangeHigh – tEvaluation Normal: - tReselectionEutra - qHyst Medium: - tReselectionEutraSfMedium - qHystSfMedium High: - tReselectionEutraSfHigh - qHystSfHigh RSRP RSRP sIntra search qHystSf Qmeas(n) Medium / qHyst qHystSf High R(n) qoffset(s) R(s) qoffset(s) Qmeas(s) tReselection Eutra sIntra search Qmeas(n) R(n) R(s) Qmeas(s) time time Cell reselection Cell tReselection reselection EutraSfMedium / tReselectionEutraSfHigh Figure 1-6: Speed Dependent Scaling of Cell Reselection 2.3 Parameter related to Idle mode mobility Some important parameters which affect idle mode mobility, is discussed below in Figure 1-7. pMax = 1000 { -30..33, 1000 } Calculates the parameter Pcompensation Value 1000 means the parameter is not sent in the system information block. cellReselectionPriority = 6 { 0..7 } The absolute priority of the carrier frequency used by the cell reselection procedure. qOffsetFreq = 0 { -24, -22, -20, -18, -16, -14, -12, -10, -8, -6, -5, -4, -3, -2, -1, 0, 1, 2, 3, 4, 5, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24 } Frequency specific offset for E-UTRAN frequencies used in connected mode. In idle mode, the negative value of this offset is used. Unit: 1 dB qOffsetCellEUtran = 0 { -24, -22, -20, -18, -16, -14, -12, -10, -8, -6, -5, -4, -3, -2, -1, 0, 1, 2, 3, 4, 5, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24 } Offset value applied to a specific cell in idle mode mobility state qQualMin = 0 { -34..-3, 0 } Value 0 means that it is not sent and UE applies in such case the (default) value of negative infinity for Qqualmin. Sent in SIB3 or SIB5.Unit: 1 dB qRxLevMin = -140 { -140..-44 } The required minimum received Reference Symbol Received Power (RSRP) level in the (E-UTRA) frequency for cell reselection. Unit: 1 dBm threshXHigh = 4 { 0..62 } The threshold used by the UE when reselecting towards the higher priority frequency X from the current serving frequency. Unit: 1 dB threshXLow = 0 { 0..62 } The threshold used in reselection towards frequency X priority from a higher priority frequency. Unit: 1 dB threshServingLow = 0 { 0..62 } Specifies the threshold that the signal strength of the serving cell must be below for cell reselection towards a lower priority Inter-Freq or IRAT Figure 1-7: Cell selection/Reselection parameters LZT1381950 R1A © Ericsson AB 2017 - 17 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop 2.3.1 Case 1: Priority Based Cell reselection example This example, taken from a live network, shows values set for parameters related to priority based cell reselection process. With the current settings, a subscriber would reselect higher priority ‘AWS band’ 1700 MHz frequency cell (i.e. Advanced Wireless Services. 1710 to 1755 MHz UL, and from 2110 to 2155 MHz DL band) over a low priority 700 MHz frequency cell. If some of the values are changed, reselection process would be impacted. Case 1 Example is shown in Figure 1-8 below, with parameters value and reselection criteria for 1700 and 700 frequency bands. Case-1: MC Priority-Based Cell Reselection Reselection Priority of AWS 1700 MHz > 700 MHz NOTE : Considering Qrxlevminoffset =0 & Pcompensation =0 hence Srxlev = Qrxlevmeas – Qrxlevmin Figure 1-8: Case-1: Cell Reselection Parameter Example 2.3.2 Case 2: 3CC CA example In this example the operator has deployed three carriers. When a network has more than one carrier, layering is important to optimize cell reselection and frequency prioritization. In this case the operator is using 2.6 GHz, 1.8 GHz, and 900 MHz LTE carriers and one WCDMA carrier. - 18 - © Ericsson AB 2017 LZT1381950 R1A LTE Mobility performance and related parameters The frequency bands of 3 carriers are B3, B7 and B8, as illustrated in Figure 1-9. Case-2: 3CC LTE-A CA : B3/B7/B8 UL + DL DL 2.6GHz High Band DL 1.8GHz High Band 2.6GHz UL + DL 1.8GHz 900MHz UL + DL 900MHz Low Band Figure 1-9: Parameter example Case-2 Different cell reselection criteria for a subscriber is explained in Figure 1-10 for a subscriber moving from high to low priority and from low to high priority frequencies. The reselection thresholds change accordingly. Priority LTE WCDM A 6 5 LTE – 2600 15M Low to high Priority (Always measure) RSRP L2600/L900 to High L1800 > QrxlevminEUTRAL1800/ (-120dBm) + Threshxhigh (8dB) = -112dBm 7 6 LTE – 1800 15M 5 5 LTE – 900 10M 4 3 UMTS - 2100 High to low Priority to L900 RSRP L1800 < QrxlevminEUTRAL1800/2600 (120dBm) + Threshservinglow (4dB) = 116dBm And RSRP L2600> QrxlevminEUTRAL2600 (120dBm) + ThreshxlowEUTRAN (8dB) = -112dBm High to low Priority to L900 RSRP L1800 < QrxlevminEUTRAL1800/2600 (120dBm) + Threshservinglow (4dB) = 116dBm And RSRP L900> QrxlevminEUTRAL900 (114dBm) + ThreshxlowEUTRAN (8dB) = -106dBm High to low Priority to WCDMA RSRP < QrxlevminEUTRAL1800/2600 (120dBm) + Threshservinglow (4dB) = -116dBm And RSCP> qRxLevMinUMTS (115dBm) + threshXLowUTRAN (6dB) = 109dBm High to low Priority to WCDMA RSRP < QrxlevminEUTRAL900 (114dBm) + Threshservinglow (4dB) = -110dBm And RSCP> qRxLevMinUMTS (115dBm) + threshXLowUTRAN (6dB) = 109dBm Figure 1-10: Case 2: 3CC idle mode LZT1381950 R1A © Ericsson AB 2017 - 19 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop 2.3.3 Case 3: Initial Traffic Balancing This case is from a customer network where a second carrier was launched, but the traffic on 2nd carrier was very low. The operator wanted to put equal traffic on the two carriers. CellreselectionPrio . made higher for 2C & Lower for 1C Figure 1-11: Case 3: Initial Traffic Balance Issue & resolution on MC site There was low/negligible traffic on the second carrier until April 21st. Faulty hardware was suspected, health check was done – but everything looked normal. On April 21st, ‘CellreselectionPriority’ parameter was optimized to influence the cell reselection process for the subscribers to achieve traffic balancing. After setting correct priorities, the second carrier started taking traffic and the two carriers shared the load- as shown in the figure above. The values for the parameter ‘cellReselectionPrio’ for the first and the second carriers were set to 4 and 5 respectively to achieve the traffic balance. 2.3.4 Case 4: Reselection parameter Example This case is another example how the idle mode mobility parameters ‘qRxLevMin’, ‘sIntraSearch’, ‘sNonIntraSearch’ could be optimized for intra and inter frequency cell reselection. A customer network is using parameter setting as shown in Figure 1-12. The impacts of using the (current) settings are described in the same table. These parameters impact cell edge coverage and intra and inter frequency search process. - 20 - © Ericsson AB 2017 LZT1381950 R1A LTE Mobility performance and related parameters Parameter Current Value Impact qRxLevMin -128 Too low and might impact at cell edge coverage during call setup sIntraSearch 62 Too early to start scanning intrafrequency-LTE, impact UE-battery-life sNonIntraSearch 48 Too early to start scanning interfrequency-LTE/inter-RAT, impact UEbattery-life Figure 1-12: Case 4: Key finding of improper parameter setting at one operator With the initial settings, the UE measurement would start at -128+ 62= -66 dBm. The proposal (after tests and field trials) was to use -124 for the ‘qRxLevMin’ and ‘sIntraSearch’ to 48 instead. This would mean the measurement would start at -124+48= -76 dBm If the UE is in good coverage in the first cell, there is no need to scan the other cell too early. This would save UE battery, but slows down the cell reselection process. Figure 1-13: Case 4: ‘qRxlevmin’ Setting discusses the impacts. Drive-test plot › ‘qRxlevmin’ consideration from operator drive-test result › The RSRP vs throughput graph suggesting small throughput at ~ -124dBm RSRP › Proposed ‘qRxLevMin’ = -124 › PROS: UE will be camped on 4 dB better coverage. › CONS: User may experience service outage in case of indoor coverage Figure 1-13: Case 4: ‘qRxlevmin’ Setting LZT1381950 R1A © Ericsson AB 2017 - 21 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop Figure 1-14: Case 4: Sintrasearch setting shows the results with new settings. Drive-test plot › ’sIntraSearch’ investigation › From field measurement, no better cell detected when 1st cell better than -66 dBm (current setting) › Proposed start intra-LTE measurement at -76 dBm ‘sIntraSearch’ = 48 › Proposed ‘sIntraSearch’ = 48 together with qRxLevMin = -124 › PROS: UE save more battery in scanning other LTE cells › CONS: slower re-selection compared to existing setting Figure 1-14: Case 4: Sintrasearch setting › Regarding triggering points for measurements and reselection: – Start to measure inter-frequency/inter-RAT – Start to re-select to inter-frequency/inter-RAT › Big gap between two triggering point ~36dB – Earlier the triggering, more the inter-frequency/inter-RAT measurements › Propose ‘sNonIntraSearch’ = 10 compared to existing 48 › PROS: UE save more battery by less scanning of inter-frequency/inter-RAT › CONS: N/A Figure 1-15: Case 4: Snonintrasearch setting - 22 - © Ericsson AB 2017 LZT1381950 R1A LTE Mobility performance and related parameters The parameter changes shorten the gap between start of measurement and cell reselection point, which in turn also saves UE battery, as shown in Figure 1-16. RSRP/RSCP Current Priority: 3 RSRP (S) qRxLevMin = -128 sIntraSearch = 62 Proposal qRxLevMin = -124 sIntraSearch = 48 Current qRxLevMin = -128 sNonIntraSearch = 48 RSRP/RSCP (N) Start search intrafreq when RSRP(s) < -66 Start search intra-freq when RSRP(s) < -76 Start search interfreq/RAT when RSRP(s) < -80 interStart search freq/RAT when RSRP(s) < -114 Priority: 2 Re-selection to inter-freq/inter-RAT, lowerpriority RSRP(s) < -116 and RSCP(n) > -107 Closer gap between start to measure and reselection point save UE battery Figure 1-16: Case 4: Summary of change Cell reselection and connected mode mobility process can also be optimized with some features which define reselection and other priorities based on subscriber, SPID, services & QCI. [These features mentioned below are discussed in detail in the course “LTE Advanced Feature”]. Multi-Layer Service-Triggered mobility feature, FAJ 121 4124 (FDD) and FAJ 221 4124 (TDD) is an enhancement of legacy feature ‘Service Triggered Mobility’. This feature overrides the Service Triggered Mobility feature when both are activated. The legacy feature ‘Mobility Control at Poor Coverage’ is a prerequisite. Subscriber Triggered Mobility, this feature enables the Radio Resource Management (RRM) and Mobility strategy in E-UTRAN to be based on user specific information by the use of SPID ‘Subscriber Profile ID for RAT/ frequency priority’. It enables individual control of mobility characteristics for a UE based on SPID. 2.4 Sticky Carrier for Idle mode mobility during IFLB The Sticky Carrier parameter setting will allow subscriber to stay on the cell where it has been transferred after load balancing. Parameter ‘incrPrioServingFreqActive’ specifies whether or not the function to set the serving frequency as sticky carrier is active in the current cell. LZT1381950 R1A © Ericsson AB 2017 - 23 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop When the Inter-Frequency Load Balancing feature is applied, consideration of the UE mobility in idle mode is highly recommended. The load balancing can thus, be improved with appropriate configuration of the UE idle mode mobility. › When the Inter-frequency Load Balancing feature is applied, consideration of the UE mobility in idle mode is recommended. › The load balancing can be improved by appropriate configuration of the UE idle mode mobility › Idle mode configuration – Sticky carrier configuration – Priority carrier configuration Figure 1-17: Idle mode behavior with IFLB functionality The sticky carrier method minimizes ping-pong mobility in the idle mode as illustrated in the Figure 1-18. › Objective: Minimize the ping-pong mobility between idle and connected mode due to IFLB › Methodology: confine UE in idle mode to the carrier frequency of the current serving cell › Setting: carrier frequency the UE is camping on has higher priority than the others Figure 1-18: Idle mode with Sticky carrier methodology - 24 - © Ericsson AB 2017 LZT1381950 R1A LTE Mobility performance and related parameters 3 EUTRA connected mode Mobility KPI The EUTRAN Mobility Success Rate KPI formula illustrated in Figure 1-19 below includes the Handover Preparation phase where resources are prepared in the target cell and the Handover Execution phase where the UE moves from the source to target cell. The KPI formula includes Intra Frequency handover where the target and source cell are on the same frequency, Inter Frequency where they are on different frequencies and IRAT handover where the target cell is in another WCDMA frequency. Figure 1-19 defines mobility KPI for preparation and execution phase. The ability to provide the requested service to the user with mobility. EUTRAN Mobility Success Rate [%]: ( = pmHoPrepSuccLteIntraF + pmHoPrepSuccLteInterF + pmHoPrepSuccWcdma pmHoPrepAttLteIntraF + pmHoPrepAttLteInterF + pmHoPrepAttWcdma ) X Equations on a cell relation pair. Refer to the CPI “Key Performance Indicator” Handover Preparation ( pmHoExeSuccLteIntraF + pmHoExeSuccLteInterF + pmHoExeSuccWcdma pmHoExeAttLteIntraF + PmHoExeAttLteInterF + pmHoExeAttWcdma ) X 100 Handover Execution Figure 1-19: EUTRAN Mobility KPI 3.1 LTE Event Review Connected mode mobility is based on measurements sent by the UE. The measurement reporting by UE can be done on an event triggered and/or on a periodic basis. LZT1381950 R1A © Ericsson AB 2017 - 25 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop The LTE events are listed in Figure 1-20. › EVENT_A1 – Serving cell becomes better than absolute threshold. › EVENT_A2 – Serving cell becomes worse than absolute threshold. › EVENT_A3 – Neighbor cell becomes amount of offset better than serving (Pcell). › EVENT_A5 – Serving cell becomes worse than absolute threshold1 AND neighbor cell becomes better than another absolute threshold2. › EVENT_A6 – Neighbor cell becomes amount of offset better than SCell › EVENT_B1 – IRAT neighbour becomes better than threshold › EVENT_B2 – Serving becomes worse than threshold1 and IRAT neighbour becomes better than threshold2 Figure 1-20: Review: LTE Events When a UE in RRC_CONNECTED mode measures poor coverage in the current LTE frequency, it informs the network by a measurement report for event A2 (serving cell becomes worse than threshold). Depending on the active features and the network configuration, the UE can be ordered to start new measurements before a handover, or a Release with Redirect (session continuity) is triggered. In case the serving cell is fully covered by another cell, the eNodeB can order the UE to perform a blind handover. [The details are explained in the ‘LTE Protocols and Procedures’ and the ‘LTE Radio Network Functionality’ courses.] The important basic features need to be activated and optimized and which are associated with the described mobility functionality are: • Coverage-Triggered Inter-Frequency Session Continuity • Coverage-Triggered GERAN Session Continuity • Coverage-Triggered WCDMA Session Continuity • Coverage-Triggered CDMA-eHRPD Session Continuity • Coverage-Triggered TD-SCDMA Session Continuity • Coverage-Triggered Inter-Frequency Handover • Coverage-Triggered WCDMA IRAT Handover • Intra-LTE Inter-Mode Handover • Coverage-Triggered TD-SCDMA IRAT Handover - 26 - © Ericsson AB 2017 LZT1381950 R1A LTE Mobility performance and related parameters A summary of coverage triggered mobility (IRAT, inter-frequency and intermode) is reviewed & illustrated in Figure 1-21 below, if MCPC feature is not activated. Session continuity Blind release with redirect (to one of the candidate freq) Only if QCI≠/1 FALSE Release with redirect (to freq ueMeasurementActive=? reported by A5/B2) a5B2MobilityTimer Bad Coverage Event A2 Serving cell worse than threshold NO Event A3orA5/B2 RBS determines a set of candidate frequenciesIs there an IF/ WCDMA/ IRAT/ Inter mode cell that fully covers the source cell? YES TRUE Event A1 Serving worse than threshold1 AND f2 / IRAT/ inter mode neighbor better than threshold2 Good coverage detected covTriggeredBlindHoAllowed=true mobilityAction=HANDOVER coverageIndicator=covers isHoAllowed=true REDIRECT mobilityAction= ? HANDOVER EUtranCellRelationTDD (E)UtranCellRelation externalUtranCellFDD lac≠0 and lac≠-1 isHoAllowed=true IF/IRAT/Inter mode HO Handover Blind IF/IRAT/Inter mode HO Figure 1-21: Review: Coverage triggered mobility 3.1.1 Mobility with “Mobility Control at Poor Coverage” feature The “flow diagram” for the ‘Mobility Control at Poor Coverage’ (MCPC) triggered functionality is shown Figure 1-22. LZT1381950 R1A © Ericsson AB 2017 - 27 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop Assuming that the license is installed and the feature is activated (with the parameter ‘featureStateMobCtrlAtPoorCov’). › Licence – featureStateMobCtrlAtPoorCov Session continuity (no QCI1) FALSE › Redirection info – redirect prio (connectedModeMobilityPrio/VoicePrio or SPID) Event A2Critical Blind release with redirect (to one of the candidate freq) Release with redirect (to freq coverageIndicator=none NO ueMeasurementActive=? reported by B2) a5B2MobilityTimer Event A2 Bad coverage detected RBS determines a set of candidate frequencies* Is there a WCDMA TRUE frequency cell that fully covers the source cell? YES ueMeasurementActive= ueMeasurementActiveIF, ueMeasurementActiveUTRAN, ueMeasurementActiveGERAN, ueMeasurementActiveCDMA2000 Event A3/A5/B2 covTriggeredBlindHoAllowed=true mobilityAction=HANDOVER coverageIndicator=covers isHoAllowed=true mobilityAction= ? HANDOVER Event A1 Good coverage detected (both RSRP&RSRQ) REDIRECT Event A2Critical UtranCellRelation isHoAllowed=true externalUtranCellFDD lac≠0 and rac≠-1 IRAT/IF HO Handover Blind IRAT/IF HO Figure 1-22: Review: “Basic” Mobility Control in Poor Coverage Operation UEs in a cell will only make inter-frequency or IRAT measurements if the ‘ueMeasurementsActive’ parameter is set to ‘true’. The UE is allowed to make these measurements before the timer specified by the parameter ‘a5B2MobilityTimer’ expires. If the default setting of 0 msec is used, measurements are not made. The decision on whether to use event A3 or A5 for other LTE frequencies depends on the setting of the ‘interFreqMeasType’ parameter. Mobility related issues analysis 4 When a network is to be optimized, engineers must gather information about the problems related to KPI degradation, different cause of failures and analysis related to these failures. This section will discuss a few issues, which generally tend to degrade the Handover Success Rate (HOSR). - 28 - © Ericsson AB 2017 LZT1381950 R1A LTE Mobility performance and related parameters 4.1 Handover failure issues The handover process can be divided into three phases - HO preparation, HO execution and data forwarding/transport network. Accordingly, optimization is based on these three issues. Three major groups of problems: › Handover preparation failures › Handover execution failures › Data forwarding failures Reasons for poor mobility include but are not limited to › Poor radio conditions › Badly tuned handover parameters Figure 1-23: Mobility Issue Analysis The reasons for poor mobility can be poor radio (RF) conditions. Different RF conditions can be analyzed on following basis: LZT1381950 R1A • Poor RF conditions/Weak coverage: Check coverage holes, overshooting, poor indoor planning issue. • Target has high uplink interference: Check performance counter indicating the uplink signal quality. Check imbalance between UL and DL quality/coverage. • Low DL quality: Check whether the transmit power of the RRU and UE falls within link budgets. • ANR PCI Conflict (Collision & Confusion): With ANR this may be a common issue, but with ANR enhancement features problematic cells can be identified. • Target exceeds cell range: The target cell is more than 15km from the UE (or the current cell range setting). In that case Ho-Exec fail occur. Need to change cell range or ‘cellindividualoffset’ between relations (as per required). © Ericsson AB 2017 - 29 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop • Target is a sleeping cell: The target cell is sleeping, need to be fixed sleeping on daily basis. Need to check the hardware and other alarms before analyzing Handover Failures due to Poor Radio Conditions › Poor RF conditions/Weak coverage › Target has high uplink interference › Imbalance between UL and DL quality › Low DL quality Check › ANR PCI Conflict (Collision & Confusion) › Target exceeds cell range › Target is a sleeping cell Figure 1-24: Mobility issue analysis HO Prep & Exec Failures 4.2 Handover prep failure- Neighbor Cell load issues Handover procedure and performance can be analyzed on preparation and execution basis. HO preparation failure can be due to neighbor cell load issue counters ‘pmHoPrepRejInHighLoad’ (MP load control), ‘pmHoPrepRejInOverload’ (procedure latency supervision) should be observed. Figure 1-25 shows the different counters to be checked for Intra/Inter HO Preparation failure. Handover Preparation Failure, Check following counters? › Handover Preparation Failure, › pmHoPrepAttLTEIntraF › pmHoPrepSuccLTEIntraF › pmHoPrepAttLTEInterF › pmHoPrepSuccLTEInterF › Neighbor cell load issue? › pmHoPrepRejInHighLoad (MP load control) › pmHoPrepRejInOverload (procedure latency supervision) Figure 1-25: Intra/Inter Handover Prep fail issue: possible cause - 30 - © Ericsson AB 2017 LZT1381950 R1A LTE Mobility performance and related parameters If above neighbor cell load counter is pegging, the following parameters, discussed in Figure 1-26 below, in the target cell should be analyzed and possibly tuned. Neighbor cell load issue, tune following parameters › tInactivityTimer , dlMaxRetxThreshold , ulMaxRetxThreshold , qOffsetcellEUtran , qRxLevMin › SchedulingAlgorithm , crsGain , tPollRetransmitUl (DRB & SRB) , tPollRetransmitDl (DRB & SRB) › noOfPucchCqiUsers , noOfPucchSrUsers , pdcchCfiMode Figure 1-26: Intra/Inter Handover Prep fail issue: possible cause If the neighbor cell load issue is found, it is recommended to reduce the number of HO by physical optimization, or by using the relevant features (e.g. UE Level Oscillating Handover Minimization & Automated Mobility Optimization). Note that MP capacity of DUS/Baseband is higher than DUL- therefore, hardware replacement might be considered. Physical optimization like (Tilt, azimuth, height of antenna) can be used to offload traffic from target cell. Target cell load can be optimized by using layering strategy, optimize Idle mode cell reselection, connected mode mobility & the load balancing parameters. 4.3 Handover failure, Target cell License and admission issue The counters associated with HO preparation failure due to target cell license and admission issues are discussed in Figure 1-27. › Neighbor cell License issue › pmHoPrepRejInLicMob › pmHoPrepRejInLicMultiErab › pmHoPrepRejInLicRlcUm › pmHoPrepRejInLicConnUsers › Admission reject › pmHoPrepRejInBearerAdmissionRej Figure 1-27: Intra/Inter Handover Prep fail issue: possible cause Neighbor cell License issues can be investigated with below checks: LZT1381950 R1A © Ericsson AB 2017 - 31 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop 4.4 • Audit the license state, feature state, service state & other parameters check related to Intra frequency HO feature, ‘MultiErabsPerUser’ feature and ‘RlcUm’ feature. • Check the license capacity for connected users of the target cell (CUL license), also the ‘Graceperiod’ should be noticed to exceed the licensed capacity limits for a limited period of time, Offload traffic from the target cell, capacity unit connected users, ‘gracePeriodTimeLeft’. • For admission reject issues, audit the UE admission control, the bearers (SRB & DRB) admission control, transport network admission control & emergency call prioritization. • Physical changes (Tilt, azimuth and the height of antenna) to offload traffic from target cell. The parameters for transport network bandwidth ‘dlTransNwBandwidth’, ‘ulTransNwBandwidth’ should also be checked. Handover Prep failure, other checks Other issues pertaining to handover preparations failure could exist, if there is something wrong with the target cell. • For the ‘Cell Down Auto’ following counters need to be checked pmCellDowntimeAuto’ /’pmCellDownLockAuto’. • If eNodeB cause, for configuration issues ensure proper configuration, especially for all new site which are integrated recently. For the ‘Cell Down Manually’, check the counters ‘pmCellDowntimeMan’/ ‘pmCellDownLockMan’, operation cause manually locked during network configuration issues ((eg. adding/changing/removing hardware) or manually locked during parameter change (eg. activate feature). These issues are discussed in Figure 1-28 Other checks › Availability Cell down: › Cell Down Auto pmCellDowntimeAuto/pmCellDownLockAuto › Transport Cause S1 issues pmErabRelAbnormalEnbActTnFail. Check for alarms like Service Unavailable(S1 connection failure) › Cell Down Manually pmCellDowntimeMan/pmCellDownLockMan Figure 1-28: Intra/Inter Handover Prep fail issue: possible cause - 32 - © Ericsson AB 2017 LZT1381950 R1A LTE Mobility performance and related parameters • Auto Restart: Ensure the eNodeB structure of the restart, MP Baseband or RRU (do complete eNB health checkup). Check RRU status, verify SW on node, S1 links down due to faulty interface board. • Transport Cause S1 issues, check the counter ‘pmErabRelAbnormalEnbActTnFail’, check for the alarms like ‘Service unavailable S1 connection failure’, hardware failure, ‘GeneralHwError’, ‘GeneralSwError’, Gigabit Ethernet Link Fault. Common alarms to be checked incldue ‘esource configuration disable’, ‘RRU fault’ and ‘link failure’. Check current alarms/ history of alarms and take necessary action according to the type of alarms. If there is something wrong with the target cell, then there may be the reasons for handover preparation fail. MME in pool are expected to share the same IP address. If HO-Preparation fail = 100%, it might be due to MME pool is different at source and target end. The command ‘get . termpointmme’ can be used to find out the MME pool IPs. If different, set as per market MME pool definition. Ensure ‘SpidHoWhiteList’ is active on the target, which has ‘primaryplmnReserved’ set to true. • Target cell is overloaded (High capacity): Need to offload target cell, check IFLB, IFO, IROW feature’s parameters or increase capacity of target cell • Target cell unavailable: The target cell is down / disabled; PLMN status is true but partOfSectorPower is set to maximum value. The command ‘lgd’ can be used to check the site status. Command ‘get . primaryplmn’ is used to find out the PLMN status and the command ‘get . power’ is used to find out the site power status. Few ‘get’ commands printouts are shown below. xxx_LTE> get . primaryplmn ======================================== MO Attribute Value ======================================== EUtranCellFDD=xxxx881 primaryPlmnAlarmSuppr false EUtranCellFDD=xxxx881 primaryPlmnReserved false Xxx_LTE> get . power 170512-08:06:20 10.203.41.14 16.0v ERBS_NODE_MODEL_G_1_312_COMPLETE stopfile=/tmp/16436 ============================================================ SubrackProdType=KRD901107/2_* maxPowerDissipation 2500 …………………………………………………….. EUtranCellFDD=xxxx883 preambleInitialReceivedTargetPower -110 LZT1381950 R1A © Ericsson AB 2017 - 33 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop EUtranCellFDD=xxxx883 prsPowerBoosting 2 EUtranCellFDD=xxxx883 rpsfPowerReduction 1000 EUtranCellFDD=xxxx883,MimoSleepFunction=1 sleepPowerControl 1 (RETAIN_SAME_POWER) SectorCarrier=1 configuredMaxTxPower 40000 SectorCarrier=1 csiRsPowerRatio 0 …………………………………………………………………………….. txPowerPersistentLock false SectorEquipmentFunction=S1 availableHwOutputPower 80000 SectorEquipmentFunction=S2 availableHwOutputPower 80000 SectorEquipmentFunction=S3 availableHwOutputPower 80000 ========================================================= The HO failure issues are discussed in Figure 1-29. Other issue, typically handover preparations fail if there is something wrong with the target cell. › MME pool not same same. If HO-Preparation fail = 100% › The command get . termpointmme can be used to find out the MME pool IPs. › SpidHoWhiteList is active on the target, which has primaryplmnReserved set to true. › Target cell is overloaded (High capacity). › Target cell Unavailable : partOfSectorPower is set to maximum value. The command lgd can be used to check the site status, get . primaryplmn is used to find out the PLMN status and get . power is used to find out the site power status. Figure 1-29: Intra/Inter Handover Prep fail issue: possible cause Check if the TAC has been defined on site according to network design. If HOPrep fail percentage is not uniform, while analyzing the cluster or the network, AMOS command ‘get . tac’ may be used to find out TAC status. Furthermore, the License issue/Software issues can also be checked. Other issue: › TAC not defined on site, as per network design. › License issue/Software issue. › Target cell has a fault (alarm, disabled cell, etc). › Site Configuration issue. Figure 1-30: Intra/Inter Handover Prep fail issue: possible cause - 34 - © Ericsson AB 2017 LZT1381950 R1A LTE Mobility performance and related parameters 5 Handover Execution failure issues and counters The Handover Execution failure counters are discussed in Figure 1-31. › Handover Execution Failure › pmHoExeAttLTEIntraF › pmHoExeSuccLTEIntraF › pmHoExeAttLTEInterF › pmHoExeSuccLTEInterF Figure 1-31: LTE Intra/Inter Handover Execution Failure: possible cause 5.1 Intra/Inter Handover Execution failure, possible cause The HO execution failure can be due to coverage/indoor coverage issue or PCI confliction or confusion, there may be some other reasons as well. Figure 1-32 shows associated counters and parameters related to handover execution and can be tuned. › PCI confliction or confusion › Coverage issues › pmBadCovEvalReport › pmRadioTbsPwrRestricted › pmBadCovSearchEvalReport › Parameter to be tuned › maximumTransmissionPower , partOfSectorPower › confOutputPower , crsGain , pdschTypeBGain › Indoor coverage issue (Small cell , PICO , DOT , Hetnet) › Physical changes (Tilt , Azimuth , height of antenna) Figure 1-32: LTE Intra/Inter Handover Execution Failure: possible cause LZT1381950 R1A © Ericsson AB 2017 - 35 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop 5.2 Handover Exec failure, Poor DL SINR issues related parameter The counters and parameters related to SINR and poor coverage issues are discussed in the Figure 1-33 below. What are the counter and parameters related to Poor DL SINR (CQI)? › pmRadioUeRepCqiDistr › pmRadioUeRepCqiDistr2 › Parameters to be tuned › maximumTransmissionPower , partOfSectorPower › confOutputPower , crsGain , pdschTypeBGain › Physical optimization (Tilt , Azimuth , height of antenna). Figure 1-33: LTE Intra/Inter Handover Execution Failure: possible cause 5.3 Handover Exec failure, UL RSSI issues related parameter The UL RSSI related counters and the parameters are discussed in Figure 1-34 below What are the UL RSSI related counters and other issues? › pmRadioRecInterferencePwr › pmRadioRecInterferencePwrPucch › pmRadioRecInterferencePwrPrb1~100 › pmSinrPucchDistr , pmSinrPuschDistr › Parameters to be checked › pZeroNominalPucch, pZeroNominalPusch › Check Power control parameters , › Check for Loose Connectors, faulty Antenna, VSWR alarms, faulty TMA Connections/settings › External interference. Figure 1-34: LTE Intra/Inter Handover Execution Failure: possible cause The ‘pmRadioRecInterferencePwr’ PDF counter illustrated in Figure 1-35 below gives the measured Noise and Interference Power on the PUSCH according to 3GPP technical specification 36.214. This counter can be used to identify if a cell has high uplink interference which can be the cause of poor Random Access. An example of the ‘pmRadioRecInterferencePwr’ counter for a cell with high and low uplink interference is given in Figure 1-35. - 36 - © Ericsson AB 2017 LZT1381950 R1A LTE Mobility performance and related parameters pmRadioRecInterferencePwr: 0,139,457290,272646,84684,33072,17023,11251,8475,3359,10791,1112,146,12,0,0 pmRadioRecInterferencePwr Low Interference pmRadioRecInterferencePwr: 0,0,0,0,0,0,0,0,0,0,0,0,0,0,585400,314600 High Interference The measured Noise and Interference Power on PUSCH, according to 36.214 PDF ranges: [0]: N+I <= -121 [1]: -121 < N+I <= -120 [2]: -120 < N+I <= -119 [3]: -119 < N+I <= -118 [4]: -118 < N+I <= -117 [5]: -117 < N+I <= -116 [6]: -116 < N+I <= -115 [7]: -115 < N+I <= -114 [8]: -114< N+I <= -113 [9]: -113 < N+I <= -112 [10]: -112 < N+I <= -108 [11]: -108 < N+I <= -104 [12]: -104 < N+I <= -100 [13]: -100 < N+I <= -96 [14]: -96 < N+I <= -92 [15]: -92 < N+I Unit: 1 dBm/PRB Figure 1-35: Uplink Interference In the case of the cell with the low uplink interference, most of the values are in the lower ‘bins’ whereas in the cell with the high uplink interference, they are in the higher bins. Once high uplink interference is detected on a cell, further investigation, including a possible site visit, may be required to find the cause of the interference. This could be due to an external source, hardware configuration or fault on site. 5.4 Handover Exec failure, Target cell RACH issue If Handover Failure is due to RACH failure CFRA can be used instead of CBRA. Poor Random Access Success Rate may be due to many factors, including overshooting cell, high uplink interference or poor RACH Root Sequence planning. Using the same ‘rachRootSequence’ in cells in the same area can lead to poor Random Access Success Rate as cells will be detecting RA-RNTIs from other cells. The ‘Automated RACH Root Sequence Allocation’ (FAJ 121 202) optional feature sets the ‘rachRootSequence’ for each cell based on its Physical Cell Identity (PCI) allocation, high-speed flag and cell range. Neighbor cells must have different PCIs to avoid PCI conflicts. The result of this is that RACH root sequence allocation based on PCI reduces root sequence overlap between cells. LZT1381950 R1A © Ericsson AB 2017 - 37 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop RACH issue related counters and parameters are discussed in Figure 1-36: LTE Intra/Inter Handover Execution Failure: possible cause Important counters and Parameters related to target cell RACH issues, › pmRaAttCfra , › pmRaSuccCfra › Parameters to be checked › preambleInitialReceivedTargetPower , cellrange › powerRampingStep (Hardcoded) , preambleTransMax (Hardcoded) › numberOfRA-Preambles (Hardcoded) › Tune RACH related parameters. › RF conditions (physical optimization (tilt/azimuth/antenna height) Figure 1-36: LTE Intra/Inter Handover Execution Failure: possible cause The equation illustrated in Figure 1-37 below can provide an approximate indication of the Contention Based Random Access Success Rate. It should be noted that the equation in Figure 1-37 may not be accurate since it estimates the number of detected preambles and assumes that a detected preamble is a randomaccess attempt. pmRaSuccCbra pmRaAttCbra pmRaUnassignedCfraFalse pmRaUnassignedCfraSum pmRaFailCbraMsg2Disc pmRaFailCbraMsg1DiscSched pmRaFailCbraMsg1DiscOoc The number of successfully detected RA Msg3 for CBRA The number of detected contention-based random access preambles The number of detected Contention Free Random Access preambles that are not allocated to any UE The total number of unassigned Contention Free Random Access preambles at each PRACH occasion during the reporting period. The number of CBRA preambles for which no random access response (RA Msg2) was sent due to expiration of the random access response window. The number of CBRA preambles that are discarded because maximum number of RA Msg3 are already scheduled. The number of CBRA preambles that are discarded because timing offset of CBRA preamble corresponds to a distance greater than configured cell range. Figure 1-37: HO Exec Fail: Contention Based Random Access Success Rate - 38 - © Ericsson AB 2017 LZT1381950 R1A LTE Mobility performance and related parameters 5.5 Overshooting Cell The ‘cellRange’ parameter illustrated in Figure 1-38, defines the maximum distance from the base station where a connection to a UE can be setup and/or maintained. The ‘Maximum Cell Range’ (FAJ 121 0869) optional feature is required to support cell ranges >15 km, illustrated in Figure 1-38 . ManagedElement +-ENodeBFunction +-EUtranCellFDD cellRange = 15 { 1..100 } [Unit: 1 km] cellRange X pmRaFailCbraMsg1DiscOoc Incremented for each preamble that is discarded because it was detected from a UE outside the configured cell range. The ‘Maximum Cell Range’ (FAJ 121 0869) optional feature is required to support cell ranges >15 km. Figure 1-38: HO Exec Fail: Overshooting Cell 5.6 Handover Prep/Exec failure, Other important check The other important check related to intra/inter HO issues are listed in Figure 1-39: Intra/Inter Frequency HO Optimization . Intra & Inter frequency related Important parameters: › Intra frequency LTE › cellIndividualOffsetEUtran ; a3offset ; hysteresisA3 ; timeToTriggerA3 , isHoAllowed , filterCoefficientEUtraRsrp , sMeasure. › Inter frequency LTE › a5Threshold1RSRP ; a5Threshold2RSRP ; hysteresisA5 ; timeToTriggerA5 , a1a2SearchThresholdRsrp , hysteresisA1A2SearchRsrp , searchEffortTime › timeToTriggerA1Search , timeToTriggerA2Search , a2CriticalThresholdRsrp Figure 1-39: Intra/Inter Frequency HO Optimization LZT1381950 R1A © Ericsson AB 2017 - 39 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop IRAT Handover HO optimization 6 The UE uses event A2 to report poor coverage to the eNodeB. The reception of an event A2 in the eNodeB is the trigger to start the evaluation process to decide how to move the UE to WCDMA or a second LTE frequency. If there is a WCDMA neighbor defined with the ‘coverageIndicator’ parameter set to ‘COVERS’ the eNodeB will use the ‘Coverage-Triggered WCDMA IRAT Handover’ optional feature to perform a blind handover to this cell. If blind handover is not possible the UE may be given some time to make measurements on WCDMA. If this time expires before the UE reports event A1 or B2 or if the UE was not given any time to measure the UE will be moved to WCDMA using the ‘Coverage-Triggered WCDMA Session Continuity’ optional feature. This feature uses ‘Release with Redirect’ (RwR) to move the UE to WCDMA which will produce longer break in connection than handover. If event A1 is reported this evaluation process is stopped and the eNodeB will wait for another A2 event. If the UE does find a WCDMA neighbor cell and reports event B2 the eNodeB will use ‘Coverage-Triggered WCDMA IRAT Handover’ optional feature to perform a measurement based handover to the reported cell or ‘Coverage-Triggered WCDMA Session Continuity’ to this cell if handover is not allowed. The IRAT based handover preparation phase begins when the source eNodeB receives a measurement report with event B2 (Serving worse than threshold1 and IRAT better than threshold2) or A2 (Serving cell worse than threshold). On reception of this message the RBS will increment the ‘pmHoPrepAtt’ counter. The RBS will make the handover decision based on the contents of the RRC ‘MEASUREMENT REPORT’ and the defined handover parameters. If the RBS decides to perform the handover it will send a S1 ‘HANDOVER REQUIRED’ message containing the necessary information to prepare the handover to the source MME e.g. target RNC Id. The source MME identifies the target SGSN and initiates the handover resource allocation procedure by sending ‘FORWARD RELOCATION REQUEST’ message. Once the resources are ordered the target SGSN sends a ‘RELOCATION REQUEST’ message to the target RNC. On reception of this message the target RNC performs Admission Control and Radio and Iu user plane resources are reserved for the accepted RABs. After this, the RNC acknowledge the target SGSN sending the message ‘RELOCATION REQUEST ACKNOWLEDGE’. On reception of this message the target SGSN will respond to the source MME with a ‘FORWARD RELOCATION RESPONSE’ message. - 40 - © Ericsson AB 2017 LZT1381950 R1A LTE Mobility performance and related parameters The source MME will then send a ‘S1 HANDOVER COMMAND’ message to the source RBS. On reception of this message the source RBS will release all resources for the UE and increment the ‘pmHoPrepSucc’ counter. Figure 1-40 below discussed different counters and parameters related to IRAT-HO. Handover Preparation Failure? › pmHoPrepAtt › pmHoPrepSucc › Possible parameters to tune? › a1a2SearchThresholdRsrp, hysteresisA1A2SearchRsrp, searchEffortTime, timeToTriggerA1Search , timeToTriggerA2Search, a2CriticalThresholdRsrp › b2Threshold1Rsrp, b2Threshold2RscpUtra, hysteresisB2, timeToTriggerB2, triggerQuantityB2, a5B2MobilityTimer Figure 1-40: IRAT Handover HO optimization 7 MCPC related counters and parameters If ‘Mobility Control at Poor Coverage’ feature is activated, ‘EUtranCell’ MO counters that should be monitored to track the feature’s performance include ‘pmCriticalBorderEvalReport’ and ‘pmBadCovSearchEvalReport’. Parameters to be optimized are: ManagedElement +-ENodeBFunction +-EUtranCellFDD +-UeMeasControl +-ReportConfigSearch a1a2SearchThresholdRsrp = -134 { -140...-44} Unit: 1 dB a1a2SearchThresholdRsrq = -165 { -195...-30} Unit: 0.1 dB a2CriticalThresholdRsrp = -140 { -140...-44} Unit: 1 dB a2CriticalThresholdRsrq = -195 { -195...-30} Unit: 0.1 dB hysteresisA1A2SearchRsrp = 20 {0...150} Unit: 0.1 dB hysteresisA1A2SearchRsrq = 15 {0...150} Unit: 0.1 dB hysteresisA2CriticalRsrp = 10 {0…150} Unit: 0.1 dB hysteresisA2CriticalRsrq = 10 {0…150} Unit: 0.1 dB timeToTriggerA1Search = 640 {0, 40, 64, 80, 100, 128, 160, 256, 320, 480, 512, 640, 1024, 1280, 2560, 5120} Unit: 1 ms timeToTriggerA2Critical = 40 {0, 40, 64, 80, 100, 128, 160, 256, 320, 480, 512, 640, 1024, 1280, 2560, 5120} Unit: 1 ms timeToTriggerA2Search = 40 {0, 40, 64, 80, 100, 128, 160, 256, 320, 480, 512, 640, 1024, 1280, 2560, 5120} Unit: 1 ms. LZT1381950 R1A © Ericsson AB 2017 - 41 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop The counters and parameters important for ‘MCPC’ based mobility are discussed in Figure 1-41 below. MCPC feature related counters. › pmCriticalBorderEvalReport › pmBadCovSearchEvalReport › MCPC feature related parameters. › a1a2SearchThresholdRsrp , hysteresisA1A2SearchRsrp , searchEffortTime › timeToTriggerA1Search , timeToTriggerA2Search , a2CriticalThresholdRsrp › hysteresisA2CriticalRsrp , timeToTriggerA2Critical , Figure 1-41: IRAT/Inter Frequency HO optimization IRAT/Inter Frequency Session continuity optimization is similar to HO process. LTE Inter Frequency Session Continuity › pmUeCtxtRelSCEutra Inter RAT Session continuity › pmUeCtxtRelSCWcdma Parameters Optimization › Similar like Handover Figure 1-42: IRAT/Inter Frequency Session continuity optimization Inter frequency load balancing counters and parameters 8 ‘Subscription quanta’ is weighted based on QCI and represents a generic cost of each bearer. The value for each QCI is configured with the parameter ‘qciSubscriptionQuanta’ (‘QciProfilePredefined’/’QciProfileOperatorDefined’). The cell subscription capacity represents an estimate of the total cell capacity. The value for each cell is configured with the ‘cellSubscriptionCapacity’ (‘EUtranCellFDD’/‘EUtranCellTDD’) parameter. The average subscription ratio is observed with the counters ‘pmLbSubRatioSum’ and ‘pmLbSubRatioSamp’ (‘EUtranCellFDD’/’EUtranCellTDD’). - 42 - © Ericsson AB 2017 LZT1381950 R1A LTE Mobility performance and related parameters Observed with the counter ‘pmLbMeasuredUe’ (‘EutranFreqRelation’), UEs are requested to perform these measurements regardless of blind handover possibilities. This enables the coverage conditions for load balancing to be configured more demanding than the coverage conditions for allowing blind handover for mobility purposes. The A5 event represents acceptable coverage for load balancing. The A5 event is configured with the parameters a5ThresholdRsrp and ‘hysteresisA5’ (‘ReportConfigEUtraInterFreqLb’). Observed with the counter ‘pmLbQualifiedUe’ (‘EUtranCellRelation’), reporting UEs are off-loaded until each LBA magnitude is reached. Observed with ‘pmHoPrepAttLteInterFLb’, ‘pmHoPrepSuccLteInterFLb’, ‘pmHoExeAttLteInterFLb’, ‘pmHoExeSuccLteInterFLb’ (‘EUtranCellRelation’) counters, the off-loading is realized by inter-frequency handover. If handover is not allowed for a certain cell relation off-loading is inhibited. This is controlled with the (legacy) parameter ‘isHoAllowed’ (‘EUtranCellRelation’). This means that load balancing should not be allowed for a cell relation if handover is not defined. The feature IFLB Activation Threshold is a licensed feature. The feature provides an activation threshold for the Inter Frequency Load Balancing (IFLB) feature, which can be used to reduce the number of load balancing actions taken by IFLB. The licensing MO instance name is ‘IFLBActivationThreshold’. To optimize IFLB below parameters can be tuned. The counters and associated parameters are illustrated Figure 1-43 below. › Inter-frequency load balancing › pmHoExeAttLteInterFLb › pmHoExeSuccLteInterFLb › pmHoPrepAttLteInterFLb › pmHoPrepSuccLteInterFLb Parameter to be tuned: › qciSubscriptionQuanta , lbThreshold , lbCeiling, cellSubscriptionCapacity , › a5Threshold1Rsrp , a5Threshold2Rsrp, hysteresisA5 Figure 1-43: Inter-frequency load balancing (IFLB) LZT1381950 R1A © Ericsson AB 2017 - 43 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop LTE Feature supporting Mobility KPI: 9 For mobility optimization, following features can be used are listed in Figure 144 below Figure 1-44: LTE Features for Mobility KPI The features, affecting load based mobility are listed in Figure 1-45 below. [The details of these features are discussed in ‘LTE Advanced Features’ training]. Figure 1-45: LTE Features affecting load based mobility(IFLB/IFO/IROW) - 44 - © Ericsson AB 2017 LZT1381950 R1A LTE Mobility performance and related parameters 10 Automated Mobility Optimization The default mobility parameters in eNodeB are not necessarily optimized when the operators deploy the network. The eNodeB with bad mobility parameters suffers from Too Early Handover, Too Late Handover, Handover to Wrong Cell, and Oscillating Handover. Manual configuration of these thresholds is difficult and wrong settings may lead to more drop connections and oscillating handovers. This feature automatically configures the ‘cellIndividualOffsett’ value per cell relation in order to reduce the number of drop connections and oscillating handovers. The major benefits of Automated Mobility Optimization are reduced need for manual tuning of mobility parameters and automated adjustment of cell border when traffic changes. It may also lead to reduction in number of drops during HO and also a reduced risk for unnecessary handovers (oscillating handovers). Too Early Handover, HO not completed: UE (User Equipment) experiences Handover Failure (HOF) during handover from Cell A to Cell B, and then does RRC reestablishment in the Cell A. According to the relation type, the counters ‘pmHoTooEarlyIntraF’ or ‘pmHoTooEarlyInterF’ are stepped on the relation from Cell A to Cell B when Too Early Handover is detected. Too Early Handover, HO completed: UE experiences Radio Link Failure (RLF) after successful handover from Cell A to Cell B, and then does RRC reestablishment in the Cell A. In this case, Cell A forwards ‘X2AP RLF INDICATION’ signal to Cell B, then Cell B sends ‘X2AP HANDOVER REPORT’ signal to Cell A. According to the relation type, the counters ‘pmHoTooEarlyIntraF’ or ‘pmHoTooEarlyInterF’ are stepped on the relation from Cell A to Cell B when Too Early Handover is detected. Too Late Handover, HO initialized: The UE experiences RLF in cell A and when handover is initialized from Cell A to Cell B, and then does RRC reestablishment in the Cell B. In this case, Cell B forwards ‘X2AP RLF INDICATION’ signal to Cell A. According to the relation type, the counters ‘pmHoTooLateIntraF’ or ‘pmHoTooLateInterF’ are stepped on the relation from Cell A to Cell B when Too Late Handover is detected. Too Late Handover, HO not initialized: The UE experiences RLF in cell A and then does RRC reestablishment in Cell B. In this case, Cell B forwards X2AP ‘RLF INDICATION’ signal to Cell A. According to the relation type, the counters ‘pmHoTooLateIntraF’ or ‘pmHoTooLateInterF’ are stepped on the relation from Cell A to Cell B when Too Late Handover is detected. LZT1381950 R1A © Ericsson AB 2017 - 45 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop Handover to Wrong Cell, HO not completed: UE experiences HOF or RLF during handover from Cell A to Cell B, and then does RRC reestablishment in the Cell C. In this case, Cell C forwards X2AP ‘RLF INDICATION’ signal to Cell A. According to the relation type, the counters ‘pmHoWrongCellIntraF’ or ‘pmHoWrongCellInterF’ will be issued on the relation from Cell A to Cell B, and the counters ‘pmHoWrongCellReestIntraF’ or ‘pmHoWrongCellReestInterF’ are stepped on the relation from Cell A to Cell C, when Handover to Wrong Cell is detected. Handover to Wrong cell, HO completed: UE experiences RLF after successful handover from Cell A to Cell B, and then does RRC reestablishment in the Cell C. In this case, Cell C forwards X2AP ‘RLF INDICATION’ signal to Cell B, then Cell B sends X2AP ‘HANDOVER REPORT’ signal to Cell A. According to the relation type, the counters ‘pmHoWrongCellIntraF’ or ‘pmHoWrongCellInterF’ will be issued on the relation from Cell A to Cell B, and the counters ‘pmHoWrongCellReestIntraF’ or ‘pmHoWrongCellReestInterF’ are stepped on the relation from Cell A to Cell C, when Handover to Wrong Cell is detected. 10.1 Parameter related to AMO feature The parameters and counters associated to AMO feature is discussed in Figure 147, 1-48 and 1-49. › EUtranCellRelation Managed Objects (MO) are automatically updated: › cellIndividualOffsetEUtran (CIO) › qOffsetCellEUtran (qOffset) › The operator can configure the adjustment range of the parameters by using parameters › cioUpperLimitAdjBySon and cioLowerLimitAdjBySon , on MO class EUtranCellFDD or EUtranCellTDD › hoOptStatNum, hoOptAdjThresholdAbs, hoOptAdjThresholdPerc Figure 1-46: Automated Mobility Optimization - 46 - © Ericsson AB 2017 LZT1381950 R1A LTE Mobility performance and related parameters The AMO feature uses following counters to perform corrective action & optimization, Figure 1-47 below. › Associated Counters pmHoTooEarlyIntraF pmHoTooEarlyInterF pmHoTooLateIntraF pmHoTooLateInterF pmHoWrongCellIntraF Too Early Handover occurs towards intra-frequency. Too Early Handover occurs towards an inter-frequency Too Late Handover occurs towards an intra-frequency Too Late Handover occurs towards an inter-freq cell Handover to Wrong Cell towards an intra-frequency cell. pmHoWrongCellInterF Handover to Wrong Cell towards an inter-frequency cell. pmHoWrongCellReestIntraF Handover to Wrong Cell with Re-estb to Intracell pmHoWrongCellReestInterF Handover to Wrong Cell with the RRC re-estab inte cell. pmHoOscIntraF pmHoOscInterF Ping-pong Handover towards intra-freq cell Feature UE Level Oscillating Handover Minimization Ping-pong Handover towards an inter-freq neighbor cell. Figure 1-47: Counters, events and parameters AMO LZT1381950 R1A © Ericsson AB 2017 - 47 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop The ‘Automated Mobility Optimization’ feature detects these problems and then corrects them by means of self-optimization of mobility parameters, Figure 1-48 These are the parameters for controlling the algorithm at various part of the evaluation : › hoOptStatTime – defines the duration of an evaluation cycle › hoOptStatNum – defines the minimum no. of attempts required › hoOptAdjThresholdAbs – Weighted average quality compare to last period in absolute terms › hoOptAdjThresholdPerc – Weighted average quality compare to last period in percentage terms › cioUpperLimitAdjBySon – maximum upper range to adjust › cioLowerLimitAdjBySon – maximum lower range to adjust › (adjustment is in the step of 1 db) Figure 1-48: Automated mobility optimization 10.1.1 Counters, events and parameters AMO The Automated Mobility Optimization feature detects these problems, and then corrects them by the means of self-optimization of mobility parameters. The parameters associated to AMO is discussed in Figure 1-49 and Figure 1-50. › Configurable Parameters Parameter MOM / SC Range Unit Defau lt hoOptStatTime MOM 1..327 67 1 hour 24 The operational cycle of the handover optimization function hoOptStatNum (Tho*) MOM 1..327 67 200 The minimal number of handovers required by the handover optimization function before adjusting handover parameters hoOptAdjThreshold Abs (Tabs*) MOM 0..327 67 5 The absolute threshold value for adjusted handover failure rate required to adjust handover parameters hoOptAdjThreshold Perc (Tperc*) MOM 0..100 0 0.1% 50 The percentage threshold value for adjusted handover failure rate required to adjust handover parameters cioUpperLimitAdjBy Son MOM 0..24 1 dB 4 Indicates the upper limit value of cellIndividualOffsetEUtran range that the SON function is allowed to adjust cioLowerLimitAdjBy Son MOM -24..0 1 dB -3 Indicates the lower limit value of cellIndividualOffsetEUtran range that the SON function is allowed to adjust Description * Used in the previous algorithm flow Figure 1-49: Counters, events and parameters AMO - 48 - © Ericsson AB 2017 LZT1381950 R1A LTE Mobility performance and related parameters › Mobility parameters to be optimized by eNB cellIndividualOffs etEUtran qOffsetCellEUtra n MOM / SC Range Un it MOM -24, -22, -20, -18, -16, 14, -12, -10, -8, -6, -5, 4, -3, -2, -1, 0, 1, 2, 3, 4, 5, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24 1 dB MOM -24, -22, -20, -18, -16, 14, -12, -10, -8, -6, -5, 4, -3, -2, -1, 0, 1, 2, 3, 4, 5, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24 1 dB Defa ult Description 0 Offset value for the neighbour cell. Used when UE is in connected mode 0 Offset value applied to a specific cell in idle mode mobility state The change of cellIndividualOffset or qOffset is 1db, after 6 and -6, it becomes 2db each step Figure 1-50: Counters, events and parameters AMO 10.2 Case Study 1, AMO trial test An operator faced mobility issues in the network, while the subscriber was entering in a poor coverage area followed be a late handover, because of that other KPIs, like HOSR, accessibility, eRAB success rate, DL user throughput was also poor and the call drop rate was also high. The operator implemented AMO feature in the network with the parameter settings shown in the Figure 1-51 below. With the default parameter setting parameters set 1, no significant improvements were observed, by setting parameters set 2, the mobility and other KPIs improved. ‘hoOptStatTime’ ‘the operational cycle of the handover optimization function’ and ‘hoOptStatNum’, ‘the minimal number of handovers required by the handover optimization function before adjusting handover parameters’. › Two sets of parameters will be trialed: Set hoOptStatTime hoOptStatNum 1 24 (default) 200 (default) 2 3 40 Figure 1-51: Case study 1: AMO Trial test LZT1381950 R1A © Ericsson AB 2017 - 49 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop Other parameters were defined as described in Figure 1-52. › Licensing=1,OptionalFeatureLicense=AutomatedMobilityOptimization featureState 1 › EUtranCellTDD=1 hoOptAdjThresholdAbs 5 › EUtranCellTDD=1 hoOptAdjThresholdPerc 50 › EUtranCellTDD=1 hoOptStatNum 200 › EUtranCellTDD=1 hoOptStatTime 24 › rrcConnReestActive true Update value: › EUtranCellRelation=1 cellIndividualOffsetEUtran › EUtranCellRelation=1 qOffsetCellEUtran Figure 1-52: Case Study 1: AMO Feature parameters The results are shown in the Figure 1-53 and Figure 1-54 below, from the first phase trial an improvement was observed in Mobility, Retainability and Accessibility KPI, in the test site. Figure 1-53: Case Study 1: AMO Performance Results - 50 - © Ericsson AB 2017 LZT1381950 R1A LTE Mobility performance and related parameters Figure 1-54: Case Study1: AMO Performance Results While the mobility performance is improved the target of 98% handover success is still not achieved. For that 2nd and 3rd phase of iteration with better parameter optimization was implemented. While the mobility performance is improved the target of 98% handover success is still not achieved Figure 1-55: Case Study1: AMO Performance Results LZT1381950 R1A © Ericsson AB 2017 - 51 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop Tried to find the other network areas to do the trail with Figure 1-56 parameters and monitor. Since AMO work on statistics, the counters were activated to monitor the processer load. › From the first phase trial, it is observed an improvement in both Mobility, Retainability and Accessibility in the test site. › Proposal: › Activate the counter to monitor the processer load. › Try to find the other network area to do the trail with below parameters and monitor. – FeatureLicense=AutomatedMobilityOptimization featureState 1 – EUtranCellFDD hoOptStatTime 2 – EUtranCellFDD hoOptStatNum 100 – EUtranCellFDD hoOptAdjThresholdAbs 2 – EUtranCellFDD hoOptAdjThresholdPerc 20 Figure 1-56: Case Study 1: AMO Performance Results Conclusion 11 Mobility Control at Poor Coverage The feature “Mobility Control at Poor Coverage” increases the probability that correct mobility action is performed in some specific deployment scenarios. With correct mobility action the outage time in case of bad coverage will be decreased with a deployment scenario dependent. This feature improves the Session Continuity features functionality to increase mobility handling flexibility at poor coverage. At least one of the below features required to start measurements in the UE, must be activated: Coverage-Triggered WCDMA Session Continuity, Coverage-Triggered TD-SCDMA Session Continuity, Coverage-Triggered CDMA-eHRPD Session Continuity, CoverageTriggered GERAN Session Continuity, or Coverage-Triggered Inter-Frequency Session Continuity. The feature “Mobility Control at Poor Coverage” uses a search zone for interfrequency and IRAT measurements. The search zone is entered by a UE when the A2 criteria for RSRQ and/or RSRQ are/is fulfilled (A2search). In the search zone, the UE is configured with: A1 (Good Coverage) measurements in order to leave the search zone. A2 (Critical Coverage) measurements in order to trigger blind mobility actions, handover (HO) or Release with Redirect (RwR). A1 and A2Critical use the same trigger (RSRP or RSRQ) as for A2Search. - 52 - © Ericsson AB 2017 LZT1381950 R1A LTE Mobility performance and related parameters A3/A5 inter-frequency and B2 IRAT measurements (Target Good Enough), according to the configured priorities (connectedModeMobility’/’VoicePrio’) in order to trigger possible measurement-based mobility actions, including IF/IRAT Release with Redirect or IF/IRAT Handover. If one of the A3/A5/B2 Target Good Enough measurement report arrives, a Release with redirect or handover to target cell is triggered. If UE has at least one bearer with QCI1 and handover is not possible, no blind RwR is performed and the system continues to wait for measurement reports. If ‘A2Critical’ is fulfilled, a blind Release with redirect or blind handover is triggered. If UE has at least one bearer with QCI1 no blind RwR is performed and the system continues to wait for measurement reports. If the UE has at least one bearer with QCI1, the attribute voicePrio is used instead of ‘ConnectedModeMobilityPrio’ to determine the priority of the frequencies. In this case, also all frequency relation MOs which do not have this attribute will not be measured. If the UE does not have an active voice bearer, the attribute ‘ConnectedModeMobilityPrio’ is used. If the elapsed time since the UE entered the search zone for the first time exceeds the value set in parameter ‘searchEffortTime’, the frequencies for which A3/A5/B2 measurements were configured are considered to be non-feasible targets, since no measurement report was triggered in the UE. Exclude all such frequencies from the list, unless this action leaves the list empty, in which case the list is left unmodified by this step. If a bearer with QCI=1 is set up while the UE is in search zone and no "Target Good Enough" measurements are running due to the expiry of the ‘a5B2MobilityTimer’, the ‘a5B2MobilityTimer’ and ‘Target Good Enough’ measurements are restarted! The attribute ‘voicePrio’ is used to determine the priority of the frequencies to be measured. The trigger quantity for the A3/A5/B2 measurements is set according to the on-going A1 Search event(s). If A1 search based on RSRP is on-going, then RSRP will be used as the trigger quantity for A3/A5/B2 measurements. If A1 search based on RSRQ is on-going, then RSRQ will be used as the trigger quantity for A3/A5/B2 measurements. If A1 search based on both RSRP and RSRQ are on-going, then A3/A5/B2 measurements will be started with both trigger quantities. If ANR is active, it will not configure any PCI measurements in the UE while it is in the search zone. Setting the ‘ConnectedModeMobilityPrio’ parameter to ‘-1’ excludes that frequency from mobility measurements and mobility actions. LZT1381950 R1A © Ericsson AB 2017 - 53 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop 11.1 MCPC Associated KPI and counters The list of the ‘EUtranCell’ MO counters and KPIs that may be used to monitor the performance of the ‘Mobility Control at Poor Coverage’ feature is represented in the Figure 1-57 below. Figure 1-57: MCPC: Related Special KPI 11.2 Parameters associated to MCPC Feature The measurement configuration is based on the parameters stated in the Figure 158 For the ‘hysteresis’ and ‘TimeToTrigger’ per event and measurement quantity refer to CPI. Note that before L16 only one MO attribute per ‘ReportConfig’ was used to configure the hysteresis, except for A1Search, A2Search and A2Critical measurements, where there were different hysteresis parameters for the RSRQ and RSRP based measurements. Since L16A there is the possibility to configure different hysteresis values between RSRP and RSRQ for the A2, A3-IF, A5 and ‘B2 ReportConfig’. For each ‘ReportConfig’ (‘A2OuterSearch’, A5, ‘B2Utra, B2Geran, ‘B2Cdma2000’, ‘B2Cdma20001xRtt, A3-IF). In addition, before L16A only one MO attribute per ‘ReportConfig’ was used to configure ‘timeToTrigger’, therefore the operator could not configure different timeToTrigger values between RSRP and RSRQ. Since L16A there is the possibility to configure different ‘timeToTrigger’ values between RSRP and RSRQ for the A1, A2, A3-IF, A5, B2 ‘ReportConfig’. - 54 - © Ericsson AB 2017 LZT1381950 R1A LTE Mobility performance and related parameters The Mobility control at poor coverage related parameters are states in Figure 158 below. A1A2 Search Threshold Description The RSRP threshold for the Event A1 and Event A2 RSRP Search a1a2SearchThresholdRsrp measurements. The RSRQ threshold for the Event A1 and Event A2 RSRQ Search a1a2SearchThresholdRsrq measurements. The RSRP hysteresis used by the Event A1 and Event A2 RSRP Search hysteresisA1A2SearchRsrp measurements. The RSRQ hysterisis used by the Event A1 and Event A2 RSRQ Search hysteresisA1A2SearchRsrq measurements. The time used to determine whether the measured frequencies shall be excluded searchEffortTime for mobility action at critical threshold or not. The time-to-trigger value for both the RSRP and RSRQ Event A1 Search timeToTriggerA1Search measurements. The time-to-trigger value for both the RSRP and RSRQ Event A2"Search timeToTriggerA2Search measurements. Critical Threshold Description a2CriticalThresholdRsrp The RSRP threshold for the Event A2 RSRP Critical Coverage measurements. a2CriticalThresholdRsrq The RSRQ threshold for the Event A2 RSRQ Critical Coverage measurements. The RSRP hysterisis used by the Event A2 RSRP Critical Coverage hysteresisA2CriticalRsrp measurements. The RSRQ hysterisis used by the Event A2 RSRQ Critical Coverage hysteresisA2CriticalRsrq measurements.will be replace after this feature is Activated Some of the A1/A2 parameters The time-to-trigger value for both the RSRP and RSRQ Event A2 Critical timeToTriggerA2Critical Coverage measurements. Figure 1-58: Mobility Control at Poor Coverage UEs in a cell will only make inter-frequency or IRAT measurements if the ‘ueMeasurementsActive’ parameter is set to ‘true’. The time in msec the UE is allowed to make these measurements is configured with the ‘a5B2MobilityTimer’ parameter. Using the default setting of 0 means that measurements are not made. The reception of A2 RSRQ triggers A5/A3/B2 with RSRQ as trigger quantity However in some networks IRAT HO/RwR decision based on RSRQ are not desired. The MO attribute ‘UeMeasControl::inhibitB2RsrqConfig’ can be used to inhibit the configuration of B2-RSRQ if the event A2 Search RSRQ is triggered when the UE enters the search zone. Since L15 it the following are possible: • Inhibit of B2-RSRQ when UE enters into search zone UEs in a cell will only make inter-frequency or IRAT measurements if the ‘ueMeasurementsActive’ parameter is set to ‘true’. The time in msec the UE is allowed to make these measurements is configured with the ‘a5B2MobilityTimer’ parameter. Using the default setting of 0 means that measurements are not made. The reception of A2 RSRQ triggers A5/A3/B2 with RSRQ as trigger quantity However in some networks IRAT HO/RwR decision based on RSRQ are not desired. The MO attribute ‘UeMeasControl::inhibitB2RsrqConfig’ can be used to inhibit the configuration of B2-RSRQ if the event A2 Search RSRQ is triggered when the UE enters the search zone. Since L15 it the following are possible: LZT1381950 R1A © Ericsson AB 2017 - 55 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop • Inhibit of B2-RSRQ when UE enters into search zone • Inhibit of B2-RSRQ when reconfiguring the “Target Good Enough” measurements in search zone. To activate B2-RSRQ measurement inhibition, the MO attribute ‘UeMeasControl: inhibitB2RsrqConfig’ must be set to true. There are few enhancements in L16 and L17 for MCPC feature which can be implemented to optimize the search zone, frequency priority, measurement quantity, with this feature operator can define optimize UE mobility in poor coverage and improve mobility, integrity, and retainability KPIs. Figure 1-59 illustrate the parameters. UE Measurement Active Par. Description Activates or deactivates EUTRAN inter-frequency measurements for mobility purposes when Mobility Control at Poor Coverage feature is enabled. Activates or deactivates UTRAN measurements for mobility purposes when ueMeasurementsActiveUTRAN Mobility Control at Poor Coverage feature is enabled. Activates or deactivates GERAN measurements for mobility purposes when ueMeasurementsActiveGERAN Mobility Control at Poor Coverage feature is enabled. Activates or deactivates CDMA2000 measurements for mobility purposes ueMeasurementsActiveCDMA2000 when Mobility Control at Poor Coverage feature is enabled. Feature specific parameters Description Used to activate or deactivate RSRQ measurements. For more information, EUtranCellFDD.zzzTemporary13 see Radio Network (DU-based node) or Manage Radio Network, UE Measurement (Baseband-based node). EUtranCellTDD.zzzTemporary13 Exists for MO classes: ueMeasurementsActiveIF EUtranCellFDD.zzzTemporary13 EUtranCellTDD.zzzTemporary13 Used to activate or deactivate the extended A5 "Target Good Enough" ENodeBFunction.zzzTemporary14 check. For more information, see Radio Network (DU-based node) or Manage Radio Network, UE Measurement (Baseband-based node). Disables or enables configuration of B2-RSRQ if Event A2Search-RSRQ is inhibitB2RsrqConfig triggered when the UE enters the search zone. Figure 1-59: Mobility Control at Poor Coverage - 56 - © Ericsson AB 2017 LZT1381950 R1A LTE Mobility performance and related parameters 11.3 Case 2, Post MCPC impact on KPIs The impact of MCPC feature on one of the customer network mobility KPIs is presented in Figure 1-60 below, in this example few statistics has been taken where major changes was observed post MCPC rollout. Figure 1-60 shows the counters values post implementation of MCPC feature, for the ‘CriticalBorderEvalReport’, ‘BadCovSearchEvalReport’ & Bad Coverage evaluation. ‘pmBadCovEvalReport’ counters, pegged multiple times for single UeRef, hence reduction in counts. Pre Post A1, A3, A5, B2 (WCDMA), user-inactivity Post Reaching Critical Coverage either (EUTRAN / GERAN) ➢Introduction of new counters post the implementation of MCPC roll out : CriticalBorderEvalReport & BadCovSearchEvalReport ➢Bad Coverage Eval Report counter pegged multiple times for Single UeRef, hence reduction in count. Figure 1-60: Case 2: Post MCPC evaluation L2600 Figure 1-61 shows post activation most of the ‘pmBadCovSearchEvalReport’ are followed by an Inter-Frequency handover Execution (~75%). Pre Post ➢ Post Activation most of the pmBadCovSearchEvalReport are followed by an InterFrequency Execution (~75%) Figure 1-61: Case 2: Post MCPC evaluation L2600 LZT1381950 R1A © Ericsson AB 2017 - 57 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop Figure 1-62 below shows increase in ANR activity post implementation of MCPC, hence an improvement in Inter Frequency HOSR(%). Number of inter frequency relations increased. Post Pre Pre Post ➢ Increase in ANR activity post implementation of MCPC, hence an improvement in Inter Frequency HOSR(%). Number of Inter frequency relations increased. Figure 1-62: Case 2: Post MCPC evaluation L2600 Figure 1-63 shows that the number of eRAB Drop reduced and Minutes/Drop increased post the implementation of MCPC feature due to improvement in the inter frequency HOSR (%). Impact of MCPC clouded by ANR created relations. Average Active (Min/Drop) Post Pre Sum ERAB Drops Active Post Pre ➢Number of Erab Drop reduced and Minutes/Drop increased post the implementation of MCPC feature due to improvement in Inter frequency HOSR(%) E Figure 1-63: Case 2: Post MCPC evaluation L2600 - 58 - © Ericsson AB 2017 LZT1381950 R1A LTE Mobility performance and related parameters Figure 1-64 below shows states for L1800, Improvement in inter Frequency HOSR (%) observed post the implementation. ANR optimized for inter frequency neighbors. UE context release to Eutra increased & to WCDMA decreased post the rollout. the UE spends less than 4 secs in the search zone before entering the critical coverage, a release with redirect will occur in accordance to ‘ConnectedmodeMobilityprio’ setting. i.e if Search Effort Time <4sec release towards EUTRA and if >4 Sec release towards GERAN. Unknown RAT. A2-crit (EUTRAN) Pre Post B2 (WCDMA) ➢Improvement in Inter Frequency HOSR(%) observed post the implementation on 5th Aug 2015. ANR optimized for inter frequency neighbors post the implementation. ➢UE context release to Eutra increased & Wcdma decreased post the rollout. If the UE spends less than 4 sec in the search zone before entering the critical coverage, a release with redirect will occur in accordance to ConnectedmodeMobilityprioSetting . i.e Search Effort Time <4sec release towards EUTRA, if >4 Sec release towards GERAN. Figure 1-64: Case 2: Post MCPC evaluation L1800 12 Multi-Layer Service-Triggered Mobility The L15B Multi-Layer Service-Triggered mobility feature, FAJ 121 4124 (FDD) and FAJ 221 4124 (TDD) is an enhancement of legacy feature “Service Triggered Mobility”. This feature overrides the Service Triggered Mobility feature when both are activated. The legacy feature “Mobility Control at Poor Coverage” is a prerequisite. Operators today have the tendency to deploy several frequency layers in LTE as well as in the other supported technologies, such as UTRAN, GERAN, CDMA2000 and Cdma20001xRtt. Thus, cells in an eNodeB have mobility relations to cells on several LTE frequencies as well as different RATs. Different frequency layers are characterized by different coverage, and capacity, thereby different mobility needs could be configured depending on the target frequency. This feature can be considered as an enhancement of the “Service Triggered Mobility” feature. The “Service Triggered Mobility” feature makes it possible to set different thresholds per QCI for the events A2, A1, A5 and B2. LZT1381950 R1A © Ericsson AB 2017 - 59 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop The solution also aims to support different A5 and B2 threshold offset values per target frequency relation or per QCI and per target frequency relation. So this feature adds an extra level of differentiation to the service triggered mobility, as different thresholds can be used for different frequencies. This feature overrides the Service Triggered Mobility feature when both features are activated. To summarize: With the Multi-Layer Service-Triggered Mobility feature it is possible for the operators to: Tune different services with different A1/A2 thresholds, Tune different frequency relations with different A5 or B2 threshold1 and threshold 2, Tune different frequency relations and different services with different A5 or B2 threshold1 and threshold2. With this feature the operators are also able to tune different frequency relations as well as different services with different QoS requirements with different mobility threshold. The operator can set different A5-T1, A5-T2, B2-T1 and B2-T2 threshold offsets on different QCIs and different frequency relations. Also, it can set different A1, A2 thresholds on different QCIs in the serving cell. This means that different QCIs could have different in/out-of poor coverage threshold (A1/A2), different bad coverage threshold (A5-T1/B2-T1) and different good enough threshold (A5T2/B2-T2) on different target frequency relation. See Figure 1-65. Note that the UE’s measurement threshold values are based on its QCI constellation. The details of this feature is discussed in LTE Advance Feature Training. A1A2Search, A5 and B2 Thresholds for QCI A5 and B2 Threshold per Target Frequency Relation Figure 1-65: Multi-Layer Service-Triggered Mobility - 60 - © Ericsson AB 2017 LZT1381950 R1A LTE Mobility performance and related parameters 13 Summary The participants should now be able to: 1 Analyze LTE Mobility performance and related parameters. 1.1 Explain Parameter related to Idle mode mobility. 1.2 Describe different LTE mobility KPIs & counters for X2HO, S1HO, IFHO, IRATHO. 1.3 Validate Mobility related parameter affecting different KPIs. 1.4 Evaluate the steps for optimization of these KPI 1.5 Analyze features with related parameters which improve LTE Mobility KPIs Figure 1-66: Summary of Chapter 1 LZT1381950 R1A © Ericsson AB 2017 - 61 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop Intentionally Blank - 62 - © Ericsson AB 2017 LZT1381950 R1A Different HOSR issues and improvement - Case Studies 2 Different HOSR issues and improvement Case Studies Objectives After completion of this chapter the participants will be able to: 2 2.1 2.2 2.3 2.4 Analyze KPI issues analysis, investigations and case studies for mobility Explain cases and investigation for different HOSR degradation and improvements Analyze some cases for CSFB degradation and solutions Investigate ANR related major issues and solutions Discuss different offline data and cases, for degradation analysis Figure 2-1: Objective of Chapter 2 LZT1381950 R1A © Ericsson AB 2017 - 63 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop Common Failure Scenario 1 To investigate the hand over failures, different protocol analyzers can be used, with investigation of message flow traces at RRC protocol, S1AP or X2AP protocols, handover failure stage can be identified. Few common failure messages are discussed in following examples. 1.1 X2 HO Preparation Phase If the handover is failed in preparation phase target eNodeB will not send positive ‘HandoverRequestsAcknowledge’ message, if admission failure at target end the message will be responded ‘HandoverRequestFailure’ or no response will be received by source cell, if there may be X2 link issue. The flow is illustrated in Figure 2-2. UE Src Trg CN Target eNB eNB eNB | | MeasurementReport |==>| |==>| | HandoverRequest | HandoverRequest pmHoPrepAttLteIntraF + |<==| | HandoverRequestAcknowledge | HandoverRequestAcknowledge pmHoPrepSuccLteIntraF + Source eNB › Common failure scenario: Target eNB responds with “Handover-Request-Failure” › Admission failure on target enodeB › Feature not supported in target enodeB Target eNB does not respond to Handover Request › Problem with X2 link Figure 2-2: Investigation: X2 HO - Preparation - 64 - © Ericsson AB 2017 LZT1381950 R1A Different HOSR issues and improvement - Case Studies 1.2 X2 Handover Execution Phase There can be multiple reasons for HO execution failure as, the target eNB does not receive RRC Reconfiguration Complete command from UE, UE has moved somewhere else, random access procedure got failed, PCI clash, UE is not located near target eNodeB, Path switch failure etc. All scenarios will result in the expiry of ‘tX2RelocOverall’ and ‘UeCtxtReleaseRequest’ with cause ‘tX2relocoverall’expiry. Figure 2-3 below shows message flow for the conditions, if target eNB does not receive RRC reconfiguration complete from UE or path switch failure. UE Src Trg CN eNB eNB | | rrcConfigurationReconfiguration |<==| pmHoExeAttLteIntraF + |==>| | SNStatusTransfer | |======>| | | | |==>| | | |<==| |<==| | UEContextRelease | Source eNB Target eNB SNStatusTransfer rrcConfigurationReconfigurationComplete PathSwitchRequest PathSwitchRequestAcknowledge UEContextRelease pmHoExeSuccLteIntraF + |<======| |======>| | | rrcConfigurationReconfiguration rrcConfigurationReconfigurationComplete Common failure scenario: › Target eNB does not receive RRC Reconfiguration Complete from UE › UE has moved somewhere else, Random access has failed › PCI clash – UE is not located near target enodeB › Path switch failure Both scenarios will result in the expiry of tX2RelocOverall, UeCtxtReleaseRequest with cause tx2relocoverall-expiry. Figure 2-3: Investigation: X2 HO - Execution LZT1381950 R1A © Ericsson AB 2017 - 65 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop 1.3 HO Execution Failure Example An example of HO execution messages is shown in Figure 2-4, which include RRC, X2AP and S1AP messages. The handover is failed due to radio conditions and RRC Connection Reconfiguration time out. The traces can be taken from Layer 3 analyzer. › Mobility Execution failure example: › [10:59.384] /[000100]LmCellPT5(Ft_INCOMING_HANDOVER) /vobs/erbs/node/lm/cellLmU/build/workspace/UehBearerHandlingSwU_mp750/UehBeare rHandlingC.cpp:899 TRACE7:UE_TRACE: CellId 2, RacUeRef 20156724, uehBearerHandlingC: received signal: timeout; port: timerRrc[0]; state: Idle; data: 0x0 › [10:59.384] /[000100]LmCellPT5(Ft_INCOMING_HANDOVER) /vobs/erbs/node/lm/cellLmU/build/workspace/UehBearerHandlingSwU_mp750/UehBeare rHandlingC.cpp:708 TRACE2:UE_TRACE: CellId 2, RacUeRef 20156724, RRC Connection Reconfiguration timeout Figure 2-4: Investigation: X2 HO - Execution - 66 - © Ericsson AB 2017 LZT1381950 R1A Different HOSR issues and improvement - Case Studies 1.4 Event Identification When Handover is being analyzed, event identification is an important step to detect RRC message ‘measurement report’, sent by the UE to the eNodeB. The report has a ‘measId’ field, to know ‘measId’ field, ‘rrcConnectionReconfiguration’ message should be analyzed, the message contains the event information, measured quantity and measured frequency. example is illustrated in Figure 2-5. › All mobility is triggered by an event. This event is reported by the UE in the measurement report. › To understand what “measID:1” means, look at the rrcConnectionReconfiguration msg These message can be taken from TEMS or any other protocol analyzer Figure 2-5: Identifying Events – example › criticalExtensions c1 : rrcConnectionReconfiguration-r8 : { › measConfig { › measObjectToAddModList { › { › measObjectId 1, › measObject measObjectEUTRA : { › carrierFreq 50, › allowedMeasBandwidth mbw6, › presenceAntennaPort1 TRUE, › neighCellConfig '10'B › } › …… › reportConfigToAddModList { › { › reportConfigId 1, › reportConfig reportConfigEUTRA : { › triggerType event : { › eventId eventA3 : { › a3-Offset 6, › reportOnLeave FALSE › ….. › measIdToAddModList { › { › measId 1, › measObjectId 1, › reportConfigId 1 › }, Refers to the carrierFreq 50 (earfcndl = 50, Band 2100) Refers to EUTRA eventA3 Definition of measId 1 Figure 2-6: Identifying Events - example LZT1381950 R1A © Ericsson AB 2017 - 67 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop In ‘rrcConnectionReconfiguration’ message EnodeB sends the measurement configuration command to inform the UE about event and measured frequency relation. Figure 2-6 shows about ‘reportConfigid’ and ‘measObjectid’ which refer about event and ‘earfcndl’. Case Studies 2 This Section presents some case studies, most of which are taken from real network setup and the parameters optimized for mobility KPI. Some of these parameter changes shall demonstrate some positive results and few negative impacts as well, these cases are being discussed here to understand and analyze the impact of different parameter optimization and the results shows in these case studies may vary network to network, based on other conditions. The main idea is to understand how different parameters may affect mobility KPIs. 2.1 Case 1- IRAT HO, optimum parameter setting to improve LTE drop rate Degradation in eRAB was observed for the one site in one of the customer network. The site was having retainability(mobility) failures & all 3 cells observed in worst offenders, site was located at border. From cell trace analysis of RSRP Vs distance graphs (approximate data), as shown in Figure 2-7, it looks like drops were due to poor coverage. Previous IRAT triggering point was ‘-117 dBm’, ‘A2thresholdRSRPPrim’ (-116), ‘HysteresisA2Prim’ (10), ‘B2threshold1RSRP’ (-116). ‘B2thresholdRSCPUtra’ (-115) dBm and MCPC feature was not activated. › Degradation in ERAB is observed for one Site, to one of the customer. › The site having retainability(mobility) issue & all 3 cells observed in worst offenders. Site is located at border, › From Celltrace analysis of RSRP Vs distance graphs (approximate data), It looks like drops were due to Poor Coverage, › Previous IRAT Triggering point : -117 dBm A2thresholdRSRPPrim (116), HysteresisA2Prim (10), B2threshold1RSRP(-116). › B2thresholdRSCPUtra = -115 dBm. › MCPC feature is not activated. › Note Please refer to IRAT triggering mechanism as per S/W version deployed in N/W, Figure 2-7: Case 1- IRAT HO, optimum parameter setting to improve LTE drop rate - 68 - © Ericsson AB 2017 LZT1381950 R1A Different HOSR issues and improvement - Case Studies RSRP Vs Distance chart is shown below in Figure 2-8, showing most of the samples in poor coverage <-116. › RSRP Vs Distance chart › Showing most of samples in poor coverage (< -116 dBm), Those calls should be redirected to UMTS before it drops in LTE. Figure 2-8: Case 1- IRAT HO, optimum parameter setting to improve LTE drop rate It was better to try with optimum IRAT settings, previous IRAT triggering point was -117 dBm ‘A2thresholdRSRPPrim’ (-116), ‘HysteresisA2Prim’ (10), ‘B2threshold1RSRP’ (-116). Recommended IRAT triggering point: …………. (Measurement basis) & ……... (Release with Redirect) by settings IRAT parameters so UE will be redirect to UMTS before drops in LTE, With MCPC feature mobility of LTE subscriber in poor coverage can be optimized, A2 and ‘A2Critical’ search zone parameters can be optimized. The parameter for MCPC are already discussed in chapter 1. › Previous IRAT Triggering point : -117 dBm A2thresholdRSRPPrim (116), HysteresisA2Prim (10), B2threshold1RSRP(-116) › What should be new recommended IRAT Trigger point › -113 (Measurement basis) & -119 (Release with Redirect) by settings below IRAT parameters so UE will be redirect to UMTS before drops in LTE Figure 2-9: Case 1- IRATHO, Parameter setting to improve LTE drop rate This case is an example of how mobility parameters are impacting mobility and retainability KPIs. LZT1381950 R1A © Ericsson AB 2017 - 69 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop By optimizing the parameters, improvements were seen as shown in Figure 2-10 › Pre-Post Retainability trend, Drops rate improved while there was no degradation in UMTS KPIs. Figure 2-10: Case 1- IRAT HO optimum parameter setting to improve LTE drop rate 2.2 Case 2: Intra frequency Oscillating Handover 2.2.1 Network issue In an operator network HO oscillation rate was high for intra frequency (ping pong handover) in a cell relation, the counter checked ‘PmHoOscIntraF’, was pegging high. Parameter optimization need to be done to improve handover success rate and to reduce HO oscillation rate, as stated in Figure 2-11: Case 2: Oscillating Handover.. › In a Network intra frequency HO oscillation rate (ping pong handover) for an area was high. › Optimization need to be done to Improve Handover success rate and to reduce Ho oscillation rate › Counter Checked: › ‘PmHoOscIntraF’ › Execution Selection Criteria : All site Figure 2-11: Case 2: Oscillating Handover - 70 - © Ericsson AB 2017 LZT1381950 R1A Different HOSR issues and improvement - Case Studies 2.2.2 Case 2: Intra frequency HO Event review Figure 2-12 is a review for A3 event and associated parameters. neighbour becomes amount offset better than serving To activate the function bestCellReleaseActive = TRUE RSRP UE measure neighbouring cells measurement reports include RSRP and RSRQ Serving LTE Cell S E S E sMeasure A3offset S E HysteresisA3 Enter Event A3 Start measuring neighbors End measuring neighbors Leave Event A3 Measurement Reports reportIntervalA3 timeToTriggerA3 reportAmountA3 (0 = continual during event) Neighbour LTE Cell Example: •A3offset: 3dBm •HysteresisA3: 1dBm •reportIntervalA3: 2s •timeToTriggerA3: 320ms •sMeasure: 0 - Disabled Time In connected mode, eNB determines when the handover occurs based on the UE reports Figure 2-12: Case 2: Intra frequency HO 2.2.3 Parameters Tuned To check the sustainability of A3, event trigger time is increased with parameter ‘TimeToTriggerA3’ from 40 to ….... › Action Taken: › Increased timer event A3 › Parameter Changing For HO Oscillation AMO feature or UE Level Oscillating Handover Minimization feature can be used. Figure 2-13: Case 2: Oscillating Handover The ‘TimeToTriggerA3’ parameter provide additional time to UE to check sustainability of received signals before triggering the event. Automated mobility optimization feature’s parameters setting can also be used to optimize oscillating HO, too early, too late and HO to wrong cell. LZT1381950 R1A © Ericsson AB 2017 - 71 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop 2.2.4 Impact: KPI improvement After changing the ‘timetotriggerA3’ parameter, HO oscillation improved with good margin, intra frequency HO also improved as shown in the Figure 2-14: Case 2: Performance result and Figure 2-15: Case 2: Performance result. Result: HO oscillation rate improved from 22.24% to 14.57% Figure 2-14: Case 2: Performance result Intra frequency success rate improved from 98.6% to 99% Figure 2-15: Case 2: Performance result - 72 - © Ericsson AB 2017 LZT1381950 R1A Different HOSR issues and improvement - Case Studies 2.3 Case 3: HOSR degraded with AMO activation 2.3.1 Issue statement An Operator has observed, that the HOSR has been degraded after AMO feature is activated in a cluster. Since AMO feature is used to improve HOSR, this degradation indicates to a wrong parameter setting, which needs to be analyzed. › HOSR degraded after activating AMO function in the cluster. › Following is the HOSR trend in the network › What may be the reason of HOSR degradation? Figure 2-16: Case 3: With AMO HOSR degraded 2.3.2 CIO statistics observation When the issue observed, first point of check was AMO basic functionality. AMO changes CIO value on the basis of different HO statistics, after HOSR degradation CIO values were observed which was changed by AMO feature. CIO values indicate that there have been, many too early handovers in the network. The reasonable results of CIO should have almost the same percentage for negative and positive values. LZT1381950 R1A © Ericsson AB 2017 - 73 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop The CIO values are analyzed as given in Figure 2-17. › CIO collected: CIO statistics based on kget files: CIO -6 -3 -2 -1 0 1 2 3 4 Count of cellIndividualOffsetEUtran 2 20053 16105 23201 239647 2646 584 212 154 Percentate 0.00066093 6.6268126 5.322137183 7.667116099 79.19492142 0.87441012 0.1929915 0.070058558 0.050891594 › What you observe from above figure? › Negative CIOstatistics value of isthe much more than the positive CIO value that Figure 2-17: Case 3: CIO network means there are many too early handover in the network. Next step is to check which neighbour relations are impacting with these CIO › The reasonable result of CIO value should be that negative and settings, it was observed that negative CIO was defined on the relations less than positive havewealmost the the same percentage. 2 KM. And if CIO isCIO negative, can analyze impact on handover. Figure 218 indicates distance in KM versus CIO counts. With negative CIO, handover become …………. and downlink interference increase. › Pick up the relations with different CIO value and calculate the distance between relations: CIO=-3 CIO=-2 CIO=-1 Distance(km) Count of Distance Distance(km) Count of Distance Distance(km) Count of Distance [0,1) 11131 [0,1) 7658 [0,1) 8267 [1,2) 6300 [1,2) 5547 [1,2) 8568 [2,3) 1790 [2,3) 1939 [2,3) 4163 [3,4) 466 [3,4) 627 [3,4) 1515 [4,5) 95 [4,5) 120 [4,5) 344 [5,6) 31 [5,6) 51 [5,6) 145 [6 31 [6 55 [6 124 › Here most of the negative CIO value showing impact on the relations less than 2km. › It will make handover harder especially when CIO=-3, and cause Figure 2-18: Case 3: CIO & distance analysis high downlink interference. - 74 - © Ericsson AB 2017 LZT1381950 R1A Different HOSR issues and improvement - Case Studies 2.3.3 Case 3: CIO = -2 & -3 distribution in different cluster › Following is the statistics of CIO=-2 & -3 distributed in cluster Figure 2-19: Case 3: CIO=-2 & -3 distribution In the cluster 2.3.4 CIO = -3 for first tier neighbors After analyzing, it was found that many first round neighbour relations was having CIO = -3, which makes handover harder to prime neighbours. As shown in Figure 2-20. › It is unreasonable that source cell 4G18_4413416E_4 has many first round neighbor relation with CIO= -3. The cell marked with red circle is neighbor relation with CIO= -3 of cell 4G18_4413416E_4 Figure 2-20: Case 3: Example of unreasonable CIO value 2.3.5 Potential reason of unreasonable CIO value AMO might set higher CIO value if handover fails due to some reasons other than RF condition, then to return to positive CIO value is difficult as discussed in Figure 2-21. ‘TimeToTriggerA3’ is an important parameter to minimize oscillating HO. LZT1381950 R1A © Ericsson AB 2017 - 75 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop AMO re-corrective CIO mechanism is illustrated in Figure 2-21 below › Mechanism of AMO re-corrective CIO › If target cell get transmission alarms during handover process and cause handover failure, AMO feature will interpret, it caused by radio conditions and adjust CIO value to -3. Once adjusting CIO to -3 dB in cell relation, it is very hard to trigger re-adjustment of CIO since target cell RSRP level should be 7dB(A3offset: 3dBm HysteresisA3: 1dBm) larger than the source cell and need handover attempt larger than 200 (‘hoOptStatNum’=200). › At this situation, target cell becomes a high DL interference cell to the source cell and will impact source cell handover to other cells. If ‘hoOptStatNum’ can be adjusted to 100 for example at second time, CIO value may can readjust to -1 or 0. › TimeToTriggerA3: › At current configuration, ‘timeToTriggerA3’=40ms, which is very short and will cause early handover. Once CIO sets to -3, it is very hard to re-adjust to 0. Larger setting of timeToTriggerA3 (eg. 320ms or 480ms) is recommended to avoid action of corrective CIO. Figure 2-21: Case 3: Potential reason of unreasonable CIO value 2.3.6 Case 3: Trail Suggestion If AMO has bad impact on HOSR ‘cioUpperLimitAdjBySon’ and ‘cioLowerLimitAdjBySon’ parameters can be adjusted. The AMO feature gives good results if the AMO parameters are correctly tuned. › Once setting CIO=-3, it is very hard to adjust CIO to -1 or 0 by AMO function itself. CIO need be set to 0 periodically especially in network which have large quantity of transmission related alarms. › ‘timeToTriggerA3’ need be set to 320 or 480ms, which is the reasonable setting to avoid early handover. › If AMO has a bad impact on HOSR, we can degrade the consequence by adjusting ‘cioUpperLimitAdjBySon’ and ‘cioLowerLimitAdjBySon’. › For a cluster the trail can be done with different set of parameter values. Figure 2-22: Case 3: Trail suggestion - 76 - © Ericsson AB 2017 LZT1381950 R1A Different HOSR issues and improvement - Case Studies 2.4 Case 4: HOSR Execution Improvement On Distance Site CIO values indicate that there have been, many too early handovers in a network. The reasonable results of CIO should have almost the same percentage for negative and positive values. Major HO failures observed between 2 cells, Distance between two neighbors was higher but those are 1st tier neighbors. Issue was RACH success rate. During HO execution. Suggest some optimization steps. Figure 2-23: Case 4: HOSR Execution Improvement On Distance Site 2.4.1 Case 4: HOSR Execution improvement with cell range change The measures taken in the cell are shown in Figure 2-24 below. › Trial done to increase ‘cellrange’ from 14 To 23 Km on target cell *2090_7A to improve RACH success rate. › ‘MaximumCellRange’ feature should be activated & PreableFormate should be 1. › Huge improvement observed in HOSR between this relation. › Need to monitor. Other KPIs where cellrange changed(Peak UL throughput might be reduced because Preamble format 1-2 consumes more UL PRBs (Time division). › The Maximum Cell Range feature is activated by setting the “featureStateMaximumCellRange” (MO: MaximumCellRange) to 1 (ACTIVATED). Figure 2-24: Case 4: HOSR Execution improvement with cell range change LZT1381950 R1A © Ericsson AB 2017 - 77 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop HOSR execution improvement is shown in Figure 2-25 below Figure 2-25: Case 4: HOSR Execution improvement with cell range changes 2.5 Case 5: Poor IRAT success rate towards poor WCDMA cell for QCI8 In an operator network IRAT Handover success rate towards WCDMA was poor, handover execution was failing. When observed the WCDMA RSCP was poor during Handover. The cells having poor UMTS coverage being selected for measurement based IRAT and eventually dropping the call. Figure 2-26 below shows the issue statement. › IRAT Handover success rate was observed towards WCDMA, handover execution was failing. › When observed the WCDMA RSCP was poor during Handover. › Cells having poor UMTS coverage being selected for measurement based IRAT and eventually dropping the call. › Which parameter to be Tuned? Figure 2-26: Case 5: Poor IRAT success rate to poor WCDMA Cell for QCI8 - 78 - © Ericsson AB 2017 LZT1381950 R1A Different HOSR issues and improvement - Case Studies 2.5.1 Case 5: Poor IRAT parameters check ‘B2Threshold2RscpUtra’ Parameter defines the threshold which UMTS cell must meet by hysteresis for a period of ‘TimeToTriggerB2’ in order to being reported by the UE provided LTE Cell is worse than ‘B2Threshold1Rsrp’+Hysteresis for a period of ‘TimeToTriggeB2’. Default value for ‘B2Threshold2RscpUtra’ is set to -115 dBm resulting in cells having poor UMTS coverage being selected for measurement based IRAT and eventually dropping the call. Changing this parameter would affect LTE as well. If the IRAT HO need to be optimized only for QCI8, QCI specific offset ‘B2Threshold2RscpUtraoffset’, a site level parameter can be optimized, so changes were made to QCI=8. Figure 2-27. › ‘B2Threshold2RscpUtra’ Parameter defines the threshold which UMTS cell must meet by hysteresis for a period of ‘timetotrigger’ in order to being reported by the UE provided LTE Cell is worse than ‘B2Threshold1Rsrp’+Hysteresis for a period of ‘timetotrigger’. › Default value for ‘B2Threshold2RscpUtra’ is set to -115 dBm resulting in cells having poor UMTS coverage being selected for measurement based IRAT and eventually dropping the call. Changing this parameter would affect LTE as well . › ‘B2Threshold2RscpUtraoffset’ is QCI specific offset site level parameter, so changes were made to QCI=8 Figure 2-27: Case 5: Poor IRAT success rate to poor WCDMA Cell for QCI8 › B2Threshold2RscpUtraoffset was changed from 0 to 10 effectively reducing B2Threshold2RscpUtra to -105dBm 2.5.2 Case 5: Poor IRAT success rate toward poor WCDMA cell for QCI8 Figure 2-28 below shows the result after ‘B2Threshold2RscpUtraoffset’ was changed from 0 to 10, effectively reducing ‘B2Threshold2RscpUtra’ to -105dBm, IRAT HO toward poor WCDMA cell reduced. Since the setting can be done per QCI basis operator can optimize these parameters on QCI level. Figure 2-28: Case 5: Poor IRAT success rate to poor WCDMA Cell for QCI8 LZT1381950 R1A © Ericsson AB 2017 - 79 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop 2.6 Poor IRAT HOSR – Other reasons & solutions Major reasons / causes behind poor IRAT handover success rate is discussed in Figure 2-29, coverage gap between LTE & 3G/Other Technology, Improper parameter settings (‘a1a2SearchThresholdRsrp’, ‘a2CriticalThresholdRSRP’, ‘b2ThresholdRscpUtra’). The Best way to optimize IRAT is by doing physical optimization. Antenna tilt can be checked through the command ‘get . electrical’. Antenna model number can be checked through the command ‘get . antenna’. Major reasons / causes behind poor IRAT handover success rate are as follows ✓ Coverage gap between LTE & 3G/Other Technology. ✓ Improper parameter setting (a1a2SearchThresholdRsrp, a2CriticalThresholdRSRP, b2ThresholdRscpUtra). The Best way to optimize IRAT is doing physical optimization. We can measure antenna tilt and adjust the EDT. › Antenna tilt can be checked through the command: get . electrical › Antenna model no. can be checked through the command get . antenna Figure 2-29: Poor IRAT HOSR – Other reasons & solutions 2.6.1 IRAT improvement Counter to be checked Following counters can be observed in order to perform IRAT analysis, ‘pmBadCovSearchEvalReport’, ‘pmCriticalBorderEvalReport’. If more samples are pegged under ‘pmBadCovSearchEvalReport’, it reflects that more subscribers have gone to the bad coverage area, consider the need to optimize MCPC feature, ‘A2 search’ and ‘A2critical’ event parameters. - 80 - © Ericsson AB 2017 LZT1381950 R1A Different HOSR issues and improvement - Case Studies One can follow the procedure as shown in Figure 2-30 below. Process to fetch the counters STEP 1: Check following two counters to perform IRAT analysis ✓pmBadCovSearchEvalReport ✓pmCriticalBorderEvalReport STEP 2: Use OSS RC and ENIQ tool to fetch these counters, Select NodeB and MO Class EUtranCellFDD, to get counter of the MO STEP 3: Select the above mentioned counters. Figure 2-30: IRAT improvement Solution Procedure Contd… A case is discussed in Figure 2-31 below, where more samples were found for the ‘pmBadCovSearchEvalReport’ counter. If more samples are found under pmBadCovSearchEvalReport. In a Network, found that in WCL01558, the gamma sector has high IRAT and also found very high samples under pmBadCovSearchEvalReport. Which Parameters to be fixed? Figure 2-31: Case 6: IRAT improvement Solution Procedure Contd… LZT1381950 R1A © Ericsson AB 2017 - 81 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop 2.6.2 Case 6: IRAT improvement with ‘a1a2SearchTresholdRsrp’ The network optimization team changed the values of the parameters ‘a1a2SearchThresholdRsrp’, ‘b2Threshold2RscpUtra’. The parameter optimized ‘a1a2SearchThresholdRsrp’ from -110 dBm to ……. dBm and ‘b2Threshold2RscpUtra’ from -115 dBm to ……... dBm on 20/04 and found good improvement in IRAT. › Change the values of the parameters › a1a2SearchThresholdRsrp, b2Threshold2RscpUtra. › Parameter optimized › a1a2SearchThresholdRsrp from -110 dBm to -112 dBm and b2Threshold2RscpUtra from -115 dBm to -105 dBm › Found good improvement in IRAT. Figure 2-32: Case 6: IRAT improvement Solution Procedure Contd… 2.6.3 High Samples for ‘pmCriticalBorderEvalReport’ High samples have been observed for the counter ‘pmCriticalBorderEvalReport’, IRAT cases also seem to be high. One sector is having high IRAT handover and high ‘pmCriticalBorderEvalReport’. Figure 2-33 shows the counter samples in the sector. If more samples are found under pmCriticalBorderEvalReport. Found SCL00197 Gamma sector has high IRAT and also found very high samples under pmCriticalBorderEvalReport. Which Parameter to be fixed? Figure 2-33: Case 6: IRAT improvement Solution Procedure Contd… - 82 - © Ericsson AB 2017 LZT1381950 R1A Different HOSR issues and improvement - Case Studies Parameters can be changed ‘a2CriticalThresholdRsrp’. The value of ‘a2CriticalThresholdRsrp’ was changed from -122 dBm to ……dBm and found good improvement in IRAT as shown in Figure 2-34 below. › We can change the value of the parameter – a2CriticalThresholdRsrp. Parameter changed: › The value of a2CriticalThresholdRsrp from -122 dBm to 124dBm › Found good improvement in IRAT. Figure 2-34: Case 6: IRAT improvement Solution Procedure Contd… 2.7 Case 7: KPIs (HOSR, Ret., Thp.) Improvement with IFHO parameter setting In a city submarket, a tunnel site has been launched with a single carrier on 1900 Band. The operator has had 2C & 3C launches on 1900 Band. Further investigation reveals, there are nearby DAS sites which are on 700 Band, and there has been no HO from 700 to 1900. › In a city submarket, Tunnel site launched with single Carrier on 1900 Band › The operator having 2C & 3C launches on 1900 Band › Issue observed, It has near by DAS sites working on 700 Band, there was no HO from 700 to 1900 band ✓ MAP view for tunnel site Figure 2-35: Case 7: KPIs Improvement with IFHO parameter setting LZT1381950 R1A © Ericsson AB 2017 - 83 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop This case is an example of an optimized parameter setting for a network having frequency layering and when subscriber is to be moved from low priority to high priority, the search parameters for Idle mode and connected mode can create impact on all KPIs. Figure 2-36 shows the parameter changed for optimization. Figure 2-37, Figure 2-38 and Figure 2-39 showing KPIs improvement after parameters optimization in the network. › Team proposed operator to changes IFHO setting & Defined these both way, resulted in all KPI Improvement › KPI Trend given in Next Slide, ( good improvement in drops rate) Figure 2-36: Case 7: KPIs (HOSR,Ret.,Thp.) Improvement with ifho parameter setting Accessibility KPI with IFHO parameter setting ➢ KPI Trend Accessibility Figure 2-37: Case 7: KPIs Improvement with IFHO parameter setting Retainability KPI improvement after IFHO parameter setting. - 84 - © Ericsson AB 2017 LZT1381950 R1A Different HOSR issues and improvement - Case Studies ➢ KPI Trend Retainability Figure 2-38: Case 7: KPIs Improvement with IFHO parameter setting PDCP Throughput improvement. ➢ KPI Trend Throughput Figure 2-39: Case 7: KPIs Improvement with IFHO parameter setting LZT1381950 R1A © Ericsson AB 2017 - 85 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop 2.8 Case 8: Coverage issue, crsGain Tuning A Network trend shows poor coverage area, there are few areas where RSRP was poorer than -105 dBm, as shown in Figure 2-40. Suggested solution was to adjust ‘CrsGain’, ‘CrsGain’ power of the resource element containing reference symbols. If the symbols contain data, then the parameter ‘pdschTypeBGain’ should be used. If we adjust the ‘CrsGain’ we should move the mobility borders according to the parameter changed, ie if ‘CrsGain’ is lowered by 3dB then the mobility thresholds should be lowered by 3dB. To optimize the network, the first step was to find the area with poor RSRP, with TEMS or any other tool the area was identified with poor coverage. as shown in Figure 2-40. Area Selection For Review: Area coverage can be plotted with TEMS tool Poor Coverage Area (RSRP ≤ 105dBm) Figure 2-40: Case 8: Coverage issue, crsGain Tuning Cell Selection Steps: Select area: RSRP < -105dBm which is consider as poor coverage area. Serving cell closes to the poor coverage area are selected to increased CRS Gain, Figure 2-41shows that area. Cell Selection Steps: 1. Select area: RSRP < -105dBm which is consider as poor coverage area. 2. Serving cell closes to the poor coverage area are selected to increased CRS Gain Figure 2-41: Case 8: Cell Selection to Increase crsGain - 86 - © Ericsson AB 2017 LZT1381950 R1A Different HOSR issues and improvement - Case Studies crsGain = 0 { -300, -200, -100, 0, 177, 300, 477, 600} sets the DL power of the cell specific reference signal (CRS) relatively a reference level defined by the power of the PDSCH type-A resource elements. If ‘crsGain’ is +3dB, the CRS power is 3dB higher than a PDSCH type A resource element. PDSCH type A resource elements are located in symbols that do not contain CRS. By Increasing CRS power coverage increased in poor coverage area. Also reduce Type-B PDSCH power accordingly. New proposed values are shown in Figure 2-42 › What is crsGain parameter and relation between CRS gain and pdschTypeBGain ? › Increasing CRS power to improve coverage in poor coverage area. › Reduce type-B PDSCH power accordingly. Parameter: MO Class Name EUtranCellFDD Parameter Name crsGain Current Network 0 New Proposed 300 ENodeBFunction pdschTypeBGain 0 1 Figure 2-42: Case 8: Coverage issue, crsGain Tuning Example Expected coverage to be improved slightly upon increased CRS power gain in poor coverage area. Some KPIs such as user throughput, RRC Connection drop could be affected due to coverage improve meaning more users are able to keep in the network. Expected coverage to be improve slightly upon increased CRS power gain in poor coverage area. Some KPIs such as user throughput, RRC Connection drop could be affected due to coverage improve meaning more users are able to keep in the network. Figure 2-43: Case 8: Coverage issue, crsGain Tuning 2.9 Case 9: Handover failed in preparation phase There may be different reasons of handover failure in preparation phase. Handover failed due to RRC setup failures in target cells. HO rejected from target cell as load is high or license issues. The case is discussed in Figure 2-44 below. LZT1381950 R1A © Ericsson AB 2017 - 87 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop › KPI Impact: Handover failed in target phase. › Root Cause: Handover failed due to RRC setup failures in target cells. HO rejected from Target cell. › Lets discuss few reasons of RRC failed in target cell? Figure 2-44: Case 9: Handover failed during RRC setup There may be different reason of RRC fail in target cell: CU license congestion, PUCCH resource congestion etc. Solution is CU license expansion, PUCCH parameter modification. After changes in ‘noOfPucchSrUsers’ & ‘noOfPucchCqiUsers’ from 160 to 320, RRC setup SR become normal and then handover recover as well. The counter to be checked in target cell are ‘pmPucchSrCqiResCongCqi’ & ‘pmPucchSrCqiResCongSr’. As shown in Figure 2-45, both parameters depend on PUCCH utilization in uplink. › There may be different reason of RRC fail in target cell: CU license congestion, PUCCH resource congestion, etc. › Solution: CU License expansion, PUCCH parameter modification. › After change noOfPucchSrUsers & noOfPucchCqiUsers from 160 to 320, RRC setup SR is normal and then handover recover as well. Figure 2-45: Case 9: Handover failed in preparation phase - 88 - © Ericsson AB 2017 LZT1381950 R1A Different HOSR issues and improvement - Case Studies 2.10 Case 10: PCI Confusion A network cell reported worst Intra LTE HO Execution, after analysis it was found due to PCI confusion between 51011-430052-6 & 51011-420211-6 cell. The relation was deleted toward 51011-430052-6. The case-10 is illustrated in Figure 2-46 below. › JK4G18_MC4424219E_6 is the worst Intra LTE HO Execution due to PCI confusion between 51011430052-6 & 51011-420211-6. Delete relation toward 51011-430052-6 has been done. Figure 2-46: Case 10: PCI Confusion The relation was again created by ANR, Relation from 4G18_MC4424219E_6 to 51011-430052-6 come again on 2016-10-26 07:08:46 created by X2. To avoid PCI conflict, the network team block the relation on 2016-10-26 10:47: 25, Figure 2-47 shows the block relation print. › Relation from 4G18_MC4424219E_6 to 51011-430052-6 come again on 2016-10-26 07:08:46 created by X2 block relation done on 201610-26 10:47:25 (UTC) Figure 2-47: Case 10: PCI Confusion LZT1381950 R1A © Ericsson AB 2017 - 89 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop 2.11 Case 11: ~100% HO Execution failures In a network cell Intra frequency HO KPI shows nearly ~ 100% failure rate, mainly towards one PCI. If so many HO fails toward one relation, one reason of failure can be PCI conflict. By checking KPI it was found HO preparation success is 100% while maximum HO execution is failing. As shown in Figure 248. KPI shows nearly ~ 100% failure rate of Intra frequency HO, mainly towards one PCI. By checking KPI it is found HO preparation Success is 100% while maximum HO execution is failing. Figure 2-48: Case 11: ~100% HO Execution failures When checked target eNB it was found that, it has same PCIs as another eNB much closer to the source. If there isn’t another eNb closer to the source, then this could indicate that the target eNb is an over shooter. In this case solution was to down tilt the target. By analyzing, Target eNB has same PCIs as another eNB much closer to the source. If there isn’t another eNb closer to the source, then this could indicate that the target eNb is an over shooter. You will then have to down tilt the target. Figure 2-49: Case 11- RRC HO Execution failures due to PCI Collision - 90 - © Ericsson AB 2017 LZT1381950 R1A Different HOSR issues and improvement - Case Studies 2.11.1 Case 11- RRC HO Execution failures due to PCI Collision Following measures can be taken for PCI collision cases, as shown in Figure 2-50 below. › To resolve ANR-PCI clashes. Delete externalenodebfunction, externaleutrancellfdd, termpttoenb and eutrancellrelation › ANR added cells from the closer eNB as neighbors in every case, resulting in a substantial increase in Handover Success Rate › Recommended to review PCI plan if two neighbor sites having same PCI value › In Below case 02215 & 02697 had same PCI, So HO failures observed in near by area. Figure 2-50: Case 11- RRC HO Execution failures due to PCI Collision 3 ANR: MOM – ANR Created ANR creates the neighboring cell relations that are needed for providing full mobility in LTE and from LTE to other technologies (WCDMA and GSM). It creates the necessary relationships between cells to succeed with handover when the user is moving in the network. The relationship is setup in time for when the handover is needed. Meaning that ANR creates the neighboring cell relation when the UE reports a new cell as best serving. ANR automatically cleans-up not used neighbors in the neighboring cell list. It means that the neighboring cell list will always be optimal. The support for ANR contains control via policies. All policies comes with default settings allowing simple activation of ANR. All changes made by ANR are displayed in OSS-RC, such as outputted in Bulk CM export. ANR shows very promising results where we believe that it will both contribute to reducing OpEx and improving network performance at the same time. The goal of ANR is to set up cellRelations to neighboring cells. To do that, ANR will create CellRelation, ExtEnb, ExtCell and TermPointToEnb. However, Frequency and FreqRelation objects are created by the operator. ANR will only search for neighboring cells on operator defined frequencies. It is bad for the UE that request a lot of measurements when you are on the cell edge. May be with many measurements dropped calls increase, because we configure measurement gaps. – That is time-gaps during which the cell will not communicate with the UE, because the UE will search for other cells. LZT1381950 R1A © Ericsson AB 2017 - 91 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop There are many other measurements that we want to configure. Load Balancing, Carrier Aggregation. Neighbor Cell Relations are defined using following MOs: • ‘EutranFreqRelation’ for Intra LTE Neighbors • ‘GERANFreqGroupRelation’ for GSM neighbors • ‘UTRANFreqRelation’ for WCDMA neighbors • ‘CDMA2000FreqBandRelation’ for CDMA2000 neighbors EnbFunction EUtranCellFDD/TDD EUtraNetwork EUtran FreqRelation EUtran CellRelation EUtran Frequency reference External EnbFunction External EUtranCellFDD/TDD TermPoint ToEnb (X2) Pre-configured MOs ANR created MOs Figure 2-51: ANR: MOM – ANR Created 3.1 Handover event triggered ANR trigger point, ANR specific thresholds should be set earlier than ordinary HO thresholds, used for increasing probability for handover over X2, used for a portion of the UEs (the rest use HO event trigger). If a surrounding cell is ranked better cell than the S-eNodeB cell (UE reports in A3 (or A5) (neighbor becomes amount of offset better than serving) for the best PCIs) but its PCI does not correspond to any of the S-eNodeB’s defined neighboring cells, the ANR function is “activated”. S-eNodeB orders the UE to retrieve the Cell Global Identity (CGI) of the best ranked surrounding cell (from now on called target eNodeB Cell, or T-eNodeB cell). The UE can only do the CGI measurement in long DRX mode (i.e. when it is inactive). The measurement can take up to 260ms hence this first handover may fail. - 92 - © Ericsson AB 2017 LZT1381950 R1A Different HOSR issues and improvement - Case Studies ANR then creates a neighbor cell relationship between the S-eNodeB cell and the T-eNodeB cell. S-eNodeB establishes SCTP with the T-eNodeB. Successful SCTP establishment confirms that the signaling transport is working. S-eNodeB requests X2 setup by sending, for each own cell, the PCI, LCI, TAC, PLMN-id and frequency. The T-eNodeB adds all S-eNodeB’s cells to its neighbor list and sends the PCI, LCI, TAC, PLMN-id and frequency of all its own cells to the S-eNodeB, which in turn add all cells to its neighbor list. Successful establishment of X2 application layer confirms that the application layer is working. The X2 interface is now established, the neighbor lists are in place and the neighbor cell relation is in place. Figure 2-52 shows the process. The S-eNodeB sends a handover request over X2 to the T-eNodeB, which in turns performs Admission Control, allocates resources and acknowledges back to S-eNodeB (with Handover Command), and S-eNodeB in turns starts handover execution by sending ‘RRC Connection Reconfiguration’ command to the UE. At the same time S-eNodeB starts Data Forwarding of unacknowledged and unsent data packets over the X2 interface to the T-eNodeB. The handover execution ends when the UE successfully accesses the T-eNodeB and sends a Handover Confirm message. The T-eNodeB notifies the MME and S-GW that the handover is completed, the bearer is switched from S-eNodeB to the T-eNodeB and the resources are release in S-eNodeB. 1. RSRP (or RSRQ) of the serving cell is below RSRP or RSRQ sMeasure 2. UE starts measuring DL RSRP (or RSRQ) 3. UE reports in A3 (neighbor becomes amount of UE starts measuring offset better than serving) for the best PCIs Serving cell 4. If the best PCI is not in the neighbor cell list the UE reports Ncell for HO (HO triggered) UE is ordered to measure and report CGI for sMeasure it 5. The UE is set for long DRX to be able to read CGI Target cell 6. UE measures and reports CGI 7. Serving eNodeB exchange cell data with target A3 eNodeB Two methods can be used to obtain the IP address: Distance from 1. over S1: eNodeBConfigTransfer and eNodeB MMEConfigTransfer (L11B and onwards) Target eNodeB 2. DNS lookup to retrieve the T-eNodeB IP address (fallback option) (L11A) Serving eNodeB 8. Serving eNodeB creates cell relationship to Target cell 9. Handover can now be executed over S1 10. X2 is setup between the eNodeBs (i.e. setup of X2 SCTP) ANR specific actions 11. Handover can be executed over X2 UENormal HO actions 12. OSS-RC is updated Figure 2-52: Handover event triggered 3.2 Periodical ANR (Inter-frequency and IRAT HO) The feature can be used together with manual optimization of neighbor lists, and is also able to remove neighbor LTE cell relations which have not been used within a particular time period (‘removeNrelTime’). LZT1381950 R1A © Ericsson AB 2017 - 93 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop For target cell, number of measurement reports is controlled for each RAT by a number of parameters e.g. ‘anrUtranMeasReportMax’, ‘anrGeranMeasReportMax’, ‘anrUtranMeasReportDecr’ , ‘anrGeranMeasReportDecr’, ‘anrUtranMeasReportIncr’, ‘anrGeranMeasReportIncr’, ‘anrUtranMeasReportMin’, ‘anrGeranMeasReportMin’. 1. A limited set of active UEs in cells are configured to measure and report ANR A3 event (neighbor becomes offset better than serving) with an ANR threshold 2. UE reports PCIs that fulfills ANR A3 threshold 3. Same procedure as for handover event triggered ANR RSRP or RSRQ UE starts measuring UE reports target cell to ANR (event triggered) Serving cell UE reports target cell for HO (HO triggered) sMeasure Target cell ANR A3 A3 Gained time ANR specific actions Normal HO actions Figure 2-53: ANR event triggered The measurement configuration sent to the UE depends on the corresponding mobility measurement configuration and are delta in relation to that. The delta values broaden the cell border and provide the Automated Neighbor Relations feature with information on new cells before they are needed for handover, providing more time to perform CGI readingAs a consequence of having smaller “A3offset” (“A3offset-a3offsetAneDelta”) the ANR A3event will happen before real A3 event creating the time to create a neighbor relation and X2 interface in time for real handover. Similar to A3, the event A5 is associated with different thresholds when it is configured for ANR, then in the handover measurement case. Notice that this is valid for FDD only. The equation for calculating the RSRP or RSRQ threshold1 and threshold 2 values for the ANR feature are listed here, RSRP: a5Threshold1Rsrp + a5Threshold1RsrpAnrDelta a5Threshold2Rsrp - a5Threshold2RsrpAnrDelta RSRQ: a5Threshold1Rsrq + a5Threshold1RsrqAnrDelta a5Threshold2Rsrq - a5Threshold2RsrqAnrDelta - 94 - © Ericsson AB 2017 LZT1381950 R1A Different HOSR issues and improvement - Case Studies After providing the eNodeB with either a specific number of reports ‘maxNoPciReportsEvent’ or a specific time ‘maxTimeEventBasedPciConf’, a UE selected for these types of measurements is then relieved from sending any more ANR-specific PCI reports during the time it is connected to the eNodeB. Figure 2-52 shows the ANR process. › › › › › › › A limited set of connected UEs are ordered to perform periodic ANR measurement (ReportStrongestCellsForSON) UEs measure and report the strongest PCI ANR assesses if reported PCIs are valid neighbors If the PCI is not in the neighbor cell list the UE is ordered to measure and report CGI for it. UEs are at the same time set in long DRX mode Neighboring cell relation is created OSS-RC is updated The updated list is used at next handover LTE WCDMA/GSM/(CDMA) ANR specific actions Normal HO actions Figure 2-54: Periodical ANR (Inter-frequency and IRAT HO) 3.3 ANR Parameter EUTRAN The minimum number of UEs that are used for NR reporting is controlled by the ‘anrUesEUtraIntraFMin’ and ‘anrUesThreshInterFMin’ parameters for intra and inter frequency neighbors respectively. Every time a neighbor cell report for handover is received with unknown neighbor cells the number of ANR reporting UEs is increased by the setting of the ‘anrUesEUtraIntraFIncrHo’ for intra frequency neighbors and ‘anrUesThreshInterFIncrHo’ for inter frequency neighbors. The maximum number ANR reporting UEs is set with the ‘anrUesEUtraIntraFMax’ and ‘anrUesThreshInterFMax’ for intra and inter frequency neighbors respectively. Every time a neighbor cell report is received with only known neighbor cells the number of ANR reporting UEs is decreased by ‘anrUesEUtraIntraFDecr’ for intra frequency neighbors and ‘anrUesThreshInterFDecr’. Figure 2-55 illustrate the important parameters which can be tuned for ANR, LZT1381950 R1A © Ericsson AB 2017 - 95 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop › PCI Reporting cellAddRsrpOffsetEutran cellAddRsrqOffsetEutran cellAddRankLimitEutran Ncell min. power offset Ncell min. quality offset Ncell highest rank anrUesEUtraIntraFMax anrUesEUtraIntraFMin anrUesEUtraIntraFDecr anrUesEUtraIntraFIncrAnr anrUesEUtraIntraFIncrHo Max. UEs with ANR thresh. Min. UEs with ANR thresh. UE decr. for report with no new PCI UE incr. for ANR report with new PCI UE incr. for HO report with new PCI › CGI Reporting cellAddRsrpThresholdEutran cellAddRsrqThresholdEutran Ncell min. received power in ANR report Ncell min. received qual. in ANR report Figure 2-55: Parameters EUTRAN 1(3) X2BlackList and X2WhiteList should be updated to set list of eNB for X2 update. › X2 Establishment x2retryTimerStart = 30s x2retryTimerMaxAuto = 1440min Start value for X2 setup retry timer Max value for X2 setup retry timer › X2 white/black listing x2BlackList x2WhiteList cannot setup X2 to listed eNB cannot remove X2 to listed eNB › HO Permission hoAllowedEutranPolicy = true If ANR shall set isHoAllowed=true or false to new Ncells Figure 2-56: Parameters 2(3) ‘removeNrelTime’, specifies how long neighbor relations without any communication to the neighbor cell must remain in the RBS. 1000 means that automatic removal is not performed. If set to 0, the MO is removed within five minutes ‘removeNcellTime’, specifies how long neighbor cells without any neighbor relations must remain in the RBS. 1000 means that automatic removal is not performed. If set to 0, the MO is removed within five minutes. ‘removeNenbTime’, specifies how long neighbor RBSs without any neighbor cells must remain in the RBS. 1000 means that automatic removal is not performed. If set to 0, the MO is removed within five minutes. - 96 - © Ericsson AB 2017 LZT1381950 R1A Different HOSR issues and improvement - Case Studies ‘isRemoveAllowed’, indicates whether the ANR function is allowed to remove this MO. It does not restrict operator removal of the MO. MOs created by ANR have this parameter set to TRUE initially. Figure 2-57 below shows some of these parameters. › › Removal of Neighbor Objects removeNrelTime = 30days › removeNcellTime = 30days › removeNenbTime = 7days › isRemoveAllowed = true EUtranCellRelation ExternalEUtranCellFDD › ctrlMode = auto ExternalENbFunction Time of inactivity before auto. Removal Time with no Nrel before auto. Removal Time with no Ncell before auto. Removal Auto removal barring Control state for ANR ExternalEUtranCellFDD EUtranCellRelation TermPointToEnb (X2) Figure 2-57: Parameters 3(3) LZT1381950 R1A © Ericsson AB 2017 - 97 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop 3.4 Observability For ANR observability OSS logs, eNodeB logs, ANR related counters and internal events can be used › OSS Log – All changes to objects made by ANR (creation/deletion/attribute settings) › Counters – pmAnrNeighbrelAdd (not active in default scanner) – pmAnrNeighbrelRem (not active in default scanner) › Events – INTERNAL_EVENT_UE_ANR_CONFIG_PCI – INTERNAL_EVENT_UE_ANR_PCI_REPORT – INTERNAL_PROC_ANR_CGI_REPORT – INTERNAL_PROC_X2_CONN_SETUP – INTERNAL_EVENT_X2_CONN_RELEASE – INTERNAL_EVENT_NEIGHBENB_CHANGE – INTERNAL_EVENT_NEIGHBCELL_CHANGE – INTERNAL_EVENT_NEIGHBREL_ADD – INTERNAL_EVENT_NEIGHBREL_REMOVE Figure 2-58: Observability 3.5 ANR Related Counters When a neighbor relation for a cell is added by the ANR function the ‘pmAnrNeighbrelAdd’ counter is incremented, the ‘pmAnrNeighbrelRem’ isincremented for all neighbor relations removed by the ANR function. - 98 - © Ericsson AB 2017 LZT1381950 R1A Different HOSR issues and improvement - Case Studies Following are the ANR related important counters which can be observed for ANR optimization. As shown in Figure 2-59. Counter pmAnrNeighbrelAdd Description Number of neighbour relations added by the ANR function. Number of neighbour relations removed by the ANR function. The counter counts the number of successfully added EUtran neighbors by the ANR function. ANR function has added a neighbor Geran cell relation. The counter is created in OSS from the event INTERNAL_EVENT_NEIGHBREL_ADD cause value Geran. ANR function has added a neighbor Utran cell relation. The counter is created in OSS from the event INTERNAL_EVENT_NEIGHBREL_ADD cause value Utran. ANR function has deleted a neighbor GERAN cell relation. The counter is created in OSS from the event INTERNAL_EVENT_ NEIGHBREL_REMOVE. ANR function has deleted a neighbor Utran cell relation. The counter is created in OSS from the event INTERNAL_EVENT_ NEIGHBREL_REMOVE. pmAnrNeighbrelRem pmAnrNeighbrelAddEUtran pmAnrNeighbrelAddGeran pmAnrNeighbrelAddUtran pmAnrNeighbrelDelGeran pmAnrNeighbrelDelUtran Figure 2-59: ANR Related Counters 3.6 Self-Organized Network & ANR Features Important SON/ANR features which can be used for Mobility KPI optimization are mentioned in Figure 2-60 below Included Functions: › Automated Neighbor Relations (Enhanced in L17.Q1) › Automated Mobility Optimization (Enhanced in L17.Q1) › Automated RACH Root Sequence Allocation › Overlaid Cell Detection › PCI Conflict Reporting › UE Level Oscillating Handover Minimization Figure 2-60: Self-Organizing Networks (FAJ 801 0435) 3.7 Automated Neighbour Relation (PCI Conflict Impact) PCI conflicts can exist in LTE Networks due to the dense deployment of cells and the limited number of PCI values. While PCI conflicts produce poor handover performance they also affect the performance of the Automatic Neighbour Relation (ANR) feature. Figure 2-61 shows an example. LZT1381950 R1A © Ericsson AB 2017 - 99 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop In this example, there is a PCI conflict between Cell B and Cell C as both are using PCI 7 on the same frequency. Cell A has Cell B in the neighbour relations list but not Cell C. Measurement reports for PCI 7 sent by UEs in Cell A will be assumed to be for cell B even if it was cell C detected by the UE. This will produce handover failures for the cell relation between cell A and cell B since cell B will always be the target cell even when cell C may have been more suitable. Furthermore, the ANR feature will not be able to add neighbour relation between cell A and cell C as is it is not using an unknown PCI. If the ANR problematic cell Policy functionality is enabled, that is the 'problematicCellPolicy' parameter is set to 'AUTO_DETECT' or 'AUTO_DETECT_AND_BAR' the 'hoSuccLevel' read only parameter will show a 'LOW' or 'MEDIUM' value in the case of PCI conflict. ManagedElement +-ENodeBFunction +-EUtranCellFDD/TDD +-EUtranFreqRelation +-EUtranCellRelation = ECGI B hoSuccLevel = LOW or MEDIUM +-EUtranCellRelation = ECGI C +-AnrFunction pciConflictDetectionEcgiMeas = true problematicCellPolicy = 1 or 2 1= AUTO_DETECT 2= AUTO_DETECT_AND_BAR pciConflictMobilityEcgiMeas = true Neighbour Relation eNodeB Cell A PCI 9 ECGI A Cell B PCI 7 ECGI B eNodeB PCI Conflict Cell C PCI 7 ECGI C PCI 7 eNodeB Measurement Report is always assumed to be for Cell B even if Cell C was detected. => Handover failures for Cell A to B relation => ANR cannot add cell relation from Cell A to C Figure 2-61: Automated Neighbour Relation (PCI Conflict Impact) 3.8 Automated Neighbour Relation (PCI Conflict Handling) The ANR function is enhanced in L17.Q1 to allow it to detect and handle PCI conflicts. Figure 2-62 shows the example. To enable ANR to detect PCI conflicts the new 'pciConflictDetectionEcgiMeas' parameter must be set to 'true' and the existing 'problematicCellPolicy' parameter set to 'AUTO_DETECT' or 'AUTO_DETECT_AND_BAR'. With these parameter settings the ANR function will request extra E-UTRAN Cell Global Identifier (ECGI) measurements for any neighbour relation where the 'hoSuccLevel' read-only parameter is showing 'LOW' or 'MEDIUM'. In this example we would get a 'LOW' or 'MEDIUM' handover success level for the EUtranCellRelation from Cell A to B because of the PCI conflict with cell C. This will mean that extra ECGI measurements will be requested whenever PCI 7 is reported. - 100 - © Ericsson AB 2017 LZT1381950 R1A Different HOSR issues and improvement - Case Studies By sending the ECGI the eNodeB will be made aware of the PCI conflict and will add neighbour relation for the conflicting cell. In this example the UE reports the ECGI of Cell C and cell relation is added. Since we have two neighbours in cell A using the same PCI and frequency we now have PCI confusion. Handover performance will be poor now for both cell relations as UEs will still only report the PCI. The 'pciConflictMobilityEcgiMeas' parameter is introduced in L17.Q1 to allow the Operator to control if ECGI measurements are requested prior to handover, in case a PCI conflict is detected. In this example setting it to true will mean that ECGI measurements will be used to support handover between Cell A and B and Cell A and C even though there is PCI confusion. ManagedElement +-ENodeBFunction +-EUtranCellFDD/TDD +-EUtranFreqRelation +-EUtranCellRelation = ECGI B hoSuccLevel = LOW or MEDIUM Added +-EUtranCellRelation = ECGI C +-AnrFunction pciConflictDetectionEcgiMeas = true New problematicCellPolicy = 1 or 2 1= AUTO_DETECT 2= AUTO_DETECT_AND_BAR pciConflictMobilityEcgiMeas = true Neighbour Relation eNodeB Cell A PCI 9 ECGI A Cell B PCI 7 ECGI B eNodeB PCI Conflict Cell C PCI 7 ECGI C ECGI C New eNodeB Extra ECGI measurements requested for any neighbour relations where ‘hoSuccLevel’ = ‘LOW’ or ‘MEDIUM’ => Allows ANR to add neighbour relation ECGI measurements used to support handover with confusion Figure 2-62: Automated Neighbour Relation (PCI Conflict Handling) 3.9 Automated Neighbour Relation (PCI Conflict Detection DRX) The new ‘AnrPciConflictDrxProfile’ Managed Object Class is introduced in L17.Q1 to allow the Operator to configure the DRX profile used by UEs to read the ECGI for ANR PCI conflict detection. The ‘anrPciConflictDrxInactivityTimer’ parameter sets the number of consecutive PDCCH subframes after successfully decoding a PDCCH that indicates an initial UL or DL user data transmission for this UE. The ‘anrPciConflictLongDrxCycle’ parameter indicates the duration of the long DRX cycle in subframes and the ‘anrPciConflictOnDurationTimer’ parameter sets the number of consecutive PDCCH subframes at DRX Cycle initialization. UEs requested to make ECGI measurements for ANR PCI Conflict Detection will follow this DRX profile.Figure 2-63 shows the example of ANR PCI conflict detection DRX. LZT1381950 R1A © Ericsson AB 2017 - 101 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop ManagedElement +-ENodeBFunction +-AnrPciConflictDrxProfile New MO anrPciConflictDrxInactivityTimer = PSF2 anrPciConflictLongDrxCycle = SF320 anrPciConflictOnDurationTimer = PSF1 eNodeB 1, 2, 3, 4, 5, 6, 8, 10, 20, 30, 40, 50, 60, 80, 100, 200, 300, 500, 750, 1280, 1920, 2560 or 0 PDCCH Sub-Frames Default = PSF2 (2 PDCCH Sub-Frames) 10, 20, 32, 40, 64, 80, 128, 160, 256, 320, 512, 640, 1024, 1280, 2048 or 2560 Sub-Frames. Default = SF320 (320 Sub-Frames) 1, 2, 3, 4, 5, 6, 8, 10, 20, 30, 40, 50, 60, 80, 100 or 200 PDCCH Sub-Frames Default = PSF1 (1 PDCCH Sub-Frame) anrPciConflictDrxInactivityTimer Receiver ON anrPciConflictLongDrxCycle ECGI Receiver OFF anrPciConflictOnDurationTime r anrPciConflictOnDurationTimer Figure 2-63: Automated Neighbour Relation (PCI Conflict Detection DRX) 3.10 Automated Mobility Optimization (Coordination with ANR) The Automated Neighbour Relation (ANR) function removes neighbour relations where no reports have been received for the time specified by the 'removeNrelTime' parameter. By changing the cell individual offset the Automated Mobility Optimization (AMO) function will impact when and if a UE moves to that cell. This can lead to the cell being no longer reported resulting in removal of the cell relation by the ANR function. Once the neighbour relation is removed the AMO function will reset the cell individual offsets for the cell and UEs will start to report it again. Once enough reports are received the neighbour relation will be created again by ANR. This creates the cycle of adding, optimizing and then later removing the relation illustrated here. Coordination between the AMO and ANR functions is introduced in L17.Q1 to avoid this cycle. If no reports are received for a neighbour relation in the time specified by the 'removeNrelTime' parameter, the AMO function is operable and a cell individual offset less than 0 is used the AMO function will reduce the absolute value of the cell individual offset by 1. This should make it more likely that UEs will report the cell and avoid the neighbour relation being removed by the ANR function. In L17.Q1 if the AMO function is 'OPERABLE' and the 'amoState' read-only parameter is set to 'ENABLED' the AMO function can prevent the ANR function from moving a 'EUtranCellRelation' Managed Object to the candidate cell relations or deleting a 'EUtranCellRelation' Managed Object when certain conditions are fulfilled. When the legacy conditions are fulfilled for moving a 'EUtranCellRelation' Managed Object to a candidate cell relation the L17.Q1 enhancement will prevent it from being moved if the cell individual offset is not equal to 0. Also when the legacy conditions are fulfilled for deleting a 'EUtranCellRelation' Managed Object the L17.Q1 enhancement will prevent it from being deleted if the cell individual offset is less than 0. This will give the AMO function more time to work on these neighbour relations. Figure 2-64. - 102 - © Ericsson AB 2017 LZT1381950 R1A Different HOSR issues and improvement - Case Studies After L17.Q1 ANR removes relation Neighbo r with no reports Bad Neighbor Cell With bad HO stat AMO adjusts Cell Individual Offset Cycle of adding, optimizing and then later removing neighbour relation Candidate list (256) ANR creates Relation Legacy conditions fulfilled for moving to Candidate list EXCEPT where Cell Individual Offset ≠ 0 When AMO function is OPERABLE and ‘amoState’ = ‘ENABLED’ Coordination between AMO and ANR avoids cycle Legacy conditions fulfilled for moving to EUtranCellRelation Delete EUtranCellRelation (128) Before L17.Q1 ANR creates Relation If Cell Individual Offset < 0 reduce absolute value by 1 Legacy conditions fulfilled for deletion EXCEPT where Cell Individual Offset < 0 Neighbo r with no reports Bad Neighbor Cell With bad HO stat AMO adjusts Cell Individual Offset Figure 2-64: Automated Mobility Optimization (Coordination with ANR) 4 CS FALLBACK CS Fallback function in EPS enables the provisioning of the CS services when the UE is served by E-UTRAN. A CS Fallback enabled terminal connected to EUTRAN can use GERAN or UTRAN to connect to the CS domain. There is also CS Fallback to CDMA2000. The parameters ‘csFallbackPrio’ for UTRAn, GERAN and CDMA2000 provide the priority for a GERAN/UTRAN/CDMA2000 frequency group at CS fallback of an emergency call. Priority levels -1 to 7 are defined where 7 is the highest priority The parameters ‘csFallbackPrioEc’ for UTRAn, GERAN and CDMA2000 provide the priority for a GERAN/UTRAN/CDMA2000 frequency group at CS fallback of an emergency call. Priority levels -1 to 7 are defined where 7 is the highest priority. The value -1 means the the frequency is excluded. The parameters ‘altCsfbTargetPrioEc’ in ‘GeranFreqGroupRelation’ MO and in ‘UtranFreqRelation’ MO provide the GERAN/UTRAN frequency group for emergency calls among all frequencies related to the cell for UEs in connected mode. Figure 2-65 shows important MO class and parameters to be tuned, for CSFB. LZT1381950 R1A © Ericsson AB 2017 - 103 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop Lte CSFB TO WCDMA WCDMA MO Class Name UtranFreqRelation UtranFreqRelation MO_Value 10713 10713 UtranFreqRelation mobilityActionCsfb CsfbToGeranUtr an featureState rimCapable OptionalFeatureLicense ExternalUtranCellFDD Parameter Name csFallbackPrio csFallbackPrioEC Recommendation 4 4 1 (RELEASE_WITH_REDIRECT_N ACC) 1 (ACTIVATED) 0 (RIM_CAPABLE) Figure 2-65: CSFB 4.1 CSFB Related Features The important CSFB related features are mentioned in Figure 2-66. Feature Number Feature Name FAJ 121 0856 CS Fallback for GSM and WCDMA LTE FAJ 121 0921 Emergency Call Handling for CS Fallback LTE FAJ 121 0876 Redirect with System Information LTE FAJ 121 0495 GERAN Session continuity, covered triggered LTE FAJ 121 0493 WCDMA Session continuity, covered triggered LTE FAJ 121 1474 LTE Cell Reselection WCDMA FAJ 121 1610 R1 CS Fallback from LTE WCDMA FAJ 121 1610 R2 CS Fallback from LTE – DMCR WCDMA FAJ 121 2179 RIM Support for System Information transfer to LTE WCDMA FAJ 121 2174 Release With Redirect to LTE WCDMA FAJ 121 155 GSM-LTE Cell Reselection GSM FAJ 121 1872 LTE to GSM NACC GSM FAJ 121 1873 Fast return to LTE after call release GSM Optional Figure 2-66: CSFB – Radio feature list - 104 - © Ericsson AB 2017 LZT1381950 R1A Different HOSR issues and improvement - Case Studies Important CSFB related KPIs are shown in Figure 2-67. An example from a network shows the values of the KPIs. CSFB call setup success rate=received Alerting times/ sponsored extended service request times*100 CSFB Call setup time: Extended service request->Alerting Reselection time back to LTE: Disconnect->Tracking area update request. Related KPI in a Project: › CSFB call setup success rate=received Alerting times/ sponsored Extended service request times*100 › CSFB Call setup time: Extended service request->Alerting › Reselection time back to LTE: Disconnect->Tracking area update request. Figure 2-67: KPI Definition Example 4.2 CSFB Cases for long CALL SETUP TIME Below Figure 2-68 shows 4 cases for longer setup time because of different reasons and CSFB failure in a project (CSFB call setup time>9.5s). Below are 4 cases for longer setup time because of different reasons and CSFB failure in Project( CSFB call setup time>9.5s): 1. Core network respond with long time 2. Cell reselection 3. Multi RRC Request 4. RSCP and Ec/No continual bad due to long distance from server site Figure 2-68: CSFB Cases for long CALL SETUP TIME 4.3 Case 1: CSFB CALL SETUP TIME due to Core Signaling According to the signaling shown in Figure 2-69, it takes about 9.8 seconds waiting for CN to response. The core network time is more than expected. Recommend to investigate Core Network side. LZT1381950 R1A © Ericsson AB 2017 - 105 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop According to the signaling it takes about 9.8 seconds waiting for CN to response. Recommend to investigate Core Network side. Figure 2-69: CSFB Case 1: Longer setup time due to Core Network signaling 4.4 CSFB failure due to long Cell reselection time According to the signaling, Figure 2-70, it takes about 5.5 seconds waiting for CN to respond. Recommend to investigate Core Network side. CSFB call setup failure due to cell reselection, which consumes nearly 7.3 seconds to get access to 3G and call proceeding, need to check reselection criteria, reselection CSFB priorities and WCDMA dominance cell in the area. According to the signaling it takes about 5.5 seconds waiting for CN to respond. Recommend to investigate Core Network side. CSFB call setup failure due to cell reselection, which consumes nearly 7.3 seconds to get access to 3G and call proceeding Start -- 11:26:22.866 7.3seconds Proceeding – 11:26:30:171 5.5 seconds CN Responding – 11:26:35:677 Figure 2-70: CSFB Case 1: CSFB failure due to long Cell reselection time 4.5 SIB reading during Cell reselection UE took 5.7 to read SIB due to cell reselection. Normally reading SIB time is 0.1-0.3s, Figure 2-71 shows the traces. - 106 - © Ericsson AB 2017 LZT1381950 R1A Different HOSR issues and improvement - Case Studies Start -- 11:26:23.171 5.7seconds to Read SIB End – 11:26:28:872 › UE took 5.7 to read SIB due to cell reselection. Normally reading SIB time is 0.1-0.3s Figure 2-71: CSFB Case 1: SIB reading during Cell reselection 4.6 Failure due to Multiple RRC connection Request CSFB call setup failure due to Multi RRC Request, which consumes nearly 6.2 seconds to get access to 3G and call proceeding, authentication in WCDMA network is also happening during CSFB process. Figure 2-72 shows the time taken. CSFB call setup failure due to Multi RRC Request, which consumes nearly 6.2 seconds to get access to 3G and call proceeding Figure 2-72: CSFB Case 1: Failure due to Multiple RRC connection Request 4.7 Failure due to Multiple RRC connection Request UE sent 3 times RRC Connection Request. This process took long time as shown in Figure 2-73. LZT1381950 R1A © Ericsson AB 2017 - 107 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop First RRC -- 12:20:18.047 2.1seconds Second RRC – 12:20:20:155 2.0seconds Third RRC – 12:20:22:154 0.4seconds RRC Completed – 12:20:22:552 › UE sent 3 times RRC CONNECTION REQUEST. This process took long time Figure 2-73: CSFB Case 1: Failure due to Multiple RRC connection Request 4.8 Failure due to Bad UTRAN Coverage It takes about 7.3s from Extend service request to call proceeding, from the message view, Ec/No continual bad due to long distance from server site. It takes about 4.1s for CN respond, recommended to investigate CN side, detail shown in CSFB Call setup time(Top 4) Figure 2-74. 1. It takes about 7.3s from Extend service request to call proceeding 2. From the message view, Ec/No continual bad due to long distance from server site. 3. It takes about 4.1s for CN respond, recommend to investigate CN side. Start 13.3s Waiting CN 4.1s End Call# EXTENDED_SERVICE_REQUEST->CM Service Request CM_Service_RequestSETUP>SETUP >Call_PROCEEDING Call_PROCEEDING>ALERTING Total 0:00:04.1 17.899 4 0:00:01.5 0:00:11.5 0:00:00.3 Figure 2-74: CSFB Case 1: Failure due to Bad UTRAN Coverage - 108 - © Ericsson AB 2017 LZT1381950 R1A Different HOSR issues and improvement - Case Studies 4.9 Case 2: UE Fail to Return to LTE After CSFB to UTRAN Problem: A customer reports that the UE fail’ to return to LTE after CSFB to UTRAN customer provide TEMS log to show that UE remained camped in UTRAN cell. Figure 2-75 shows the traces. Need to interpret the logs, inorder to analyze the issue. › Problem: customer reports that UE fail to return to LTE after CSFB to UTRAN › Customer provide TEMs log to show that UE remained camped in UTRAN cell Figure 2-75: Case 2: UE Fail to Return to LTE After CSFB to UTRAN 4.9.1 Case 2: UE Fail to Return to LTE After CSFB to UTRAN › Facts: – Successful CSFB to UTRAN Cell – Need to check if eUTRAN Cell prior is higher than UTRAN Cell OFFLINE_L1M_0655> get . cellReselectionPriority ======================================================================================================== ========= MO Attribute Value ======================================================================================================== ========= EUtranCellFDD=L1M_0655A, UtranFreqRelation=10838_TME cellReselectionPriority 2 EUtranCellFDD=L1M_0655A, EUtranFreqRelation=1691 cellReselectionPriority 6 EUtranCellFDD=L1M_0655A, UtranFreqRelation=10588 cellReselectioPriority 4 EUtranCellFDD=L1M_0655A, UtranFreqRelation=10563 cellReselectionPriority 4 EUtranCellFDD=L1M_0655A, UtranFreqRelation=10813_TME cellReselectionPriority 2 EUtranCellFDD=L1M_0655A, UtranFreqRelation=10788_TME cellReselectionPriority 2 EUtranCellFDD=L1M_0655A, UtranFreqRelation=3011_TME cellReselectionPriority 2 EUtranCellFDD=L1M_0655B, UtranFreqRelation=10838_TME cellReselectionPriority 2 EUtranCellFDD=L1M_0655B, EUtranFreqRelation=1691 cellReselectionPriority 6 EUtranCellFDD=L1M_0655B, UtranFreqRelation=10588 cellReselectionPriority 4 EUtranCellFDD=L1M_0655B, UtranFreqRelation=10563 cellReselectionPriority 4 EUtranCellFDD=L1M_0655B, UtranFreqRelation=10813_TME cellReselectionPriority 2 EUtranCellFDD=L1M_0655B, UtranFreqRelation=10788_TME cellReselectionPriority 2 EUtranCellFDD=L1M_0655B, UtranFreqRelation=3011_TME cellReselectionPriority 2 EUtranCellFDD=L1M_0655C, EUtranFreqRelation=1691 cellReselectionPriority 6 EUtranCellFDD=L1M_0655C, UtranFreqRelation=10838_TME cellReselectionPriority 2 EUtranCellFDD=L1M_0655C, UtranFreqRelation=10588 cellReselectionPriority 4 EUtranCellFDD=L1M_0655C, UtranFreqRelation=10563 cellReselectionPriority 4 EUtranCellFDD=L1M_0655C, UtranFreqRelation=10813_TME cellReselectionPriority 2 EUtranCellFDD=L1M_0655C, UtranFreqRelation=10788_TME cellReselectionPriority 2 EUtranCellFDD=L1M_0655C, UtranFreqRelation=3011_TME cellReselectionPriority 2 Figure 2-76: Case 2: UE Fail to Return to LTE After CSFB to UTRAN LZT1381950 R1A © Ericsson AB 2017 - 109 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop The UE did successful CSFB to UTRAN Cell, but fails to return back to LTE, Need to check if eUTRAN Cell priority is higher than UTRAN Cell, LTE priority should be higher than WCDMA. Figure 2-76 shows ’get’ command printout for cellreselection priority. 4.9.2 Case 2: UE Fail to Return to LTE After CSFB to UTRAN Investigation need to be performed OSS reports, UE specific trace, eNB cell trace, check configuration. Adjust eUTRAN cell priority to be higher than UTRAN cell priority. It is discussed in Figure 2-77. › Investigation: OSS reports, UE specific trace, eNB cell trace, Check Configuration › Adjust eUTRAN Cell Priority to be Higher than UTRAN Cell Priority OFFLINE_L1M_0655> get . cellReselectionPriority ================================================================================================== =============== MO Attribute Value ================================================================================================== =============== EUtranCellFDD=L1M_0655A, UtranFreqRelation=10838_TME cellReselectionPriority 2 EUtranCellFDD=L1M_0655A, EUtranFreqRelation=1691 cellReselectionPriority 6 EUtranCellFDD=L1M_0655A, UtranFreqRelation=10588 cellReselectionPriority 4 EUtranCellFDD=L1M_0655A, UtranFreqRelation=10563 cellReselectionPriority 4 EUtranCellFDD=L1M_0655A, UtranFreqRelation=10813_TME cellReselectionPriority 2 EUtranCellFDD=L1M_0655A, UtranFreqRelation=10788_TME cellReselectionPriority 2 EUtranCellFDD=L1M_0655A, UtranFreqRelation=3011_TME cellReselectionPriority 2 EUtranCellFDD=L1M_0655B, UtranFreqRelation=10838_TME cellReselectionPriority 2 EUtranCellFDD=L1M_0655B, EUtranFreqRelation=1691 cellReselectionPriority 6 EUtranCellFDD=L1M_0655B, UtranFreqRelation=10588 cellReselectionPriority 4 EUtranCellFDD=L1M_0655B, UtranFreqRelation=10563 cellReselectionPriority 4 EUtranCellFDD=L1M_0655B, UtranFreqRelation=10813_TME cellReselectionPriority 2 EUtranCellFDD=L1M_0655B, UtranFreqRelation=10788_TME cellReselectionPriority 2 EUtranCellFDD=L1M_0655B, UtranFreqRelation=3011_TME cellReselectionPriority 2 EUtranCellFDD=L1M_0655C, EUtranFreqRelation=1691 cellReselectionPriority 6 EUtranCellFDD=L1M_0655C, UtranFreqRelation=10838_TME cellReselectionPriority 2 EUtranCellFDD=L1M_0655C, UtranFreqRelation=10588 cellReselectionPriority 4 EUtranCellFDD=L1M_0655C, UtranFreqRelation=10563 cellReselectionPriority 4 EUtranCellFDD=L1M_0655C, UtranFreqRelation=10813_TME cellReselectionPriority 2 EUtranCellFDD=L1M_0655C, UtranFreqRelation=10788_TME cellReselectionPriority 2 EUtranCellFDD=L1M_0655C, UtranFreqRelation=3011_TME cellReselectionPriority 2 Figure 2-77: Case 2: UE Fail to Return to LTE After CSFB to UTRAN Answer - 110 - © Ericsson AB 2017 LZT1381950 R1A Different HOSR issues and improvement - Case Studies 4.10 Case 3: UE took more time when return back to LTE NW1 In some networks the return time to LTE after a CS call finished in WCDMA was analyzed, the KPI was not meeting, when compared with other networks. The UE took more time to return to LTE in the network 1 than the other two networks, the message trace for 3 networks is shown in Figure 2-78, Figure 2-79 & Figure 2-80 respectably. UE took more time in first network ‘NW1’ as shown in Figure 2-78 compared with other 2 networks ‘NW2’ and ‘NW3’ shown in later figures. This is an exercise for participants to see 3 traces and find the difference in message trace in the figures. Figure 2-78: Case 3: UE took more time when return back to LTE NW1 Figure 2-79: Case 3: UE took more time when return back to LTE NW2 LZT1381950 R1A © Ericsson AB 2017 - 111 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop UE took less time in network-3, Figure 2-80 below, as compare to network-1. Let’s analyze the trace. Figure 2-80: Case 3: UE took more time when return back to LTE NW3 - 112 - © Ericsson AB 2017 LZT1381950 R1A Different HOSR issues and improvement - Case Studies 5 Summary The participants should now be able to: 2 2.1 2.2 2.3 2.4 Analyze KPI issues analysis, investigations and case studies for mobility Explain cases and investigation for different HOSR degradation and improvements Analyze some cases for CSFB degradation and solutions Investigate ANR related major issues and solutions Discuss different offline data and cases, for degradation analysis Figure 2-81: Summary of Chapter 2 LZT1381950 R1A © Ericsson AB 2017 - 113 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop Intentionally Blank - 114 - © Ericsson AB 2017 LZT1381950 R1A LTE Integrity KPIs Performance and related Parameters 3 LTE Integrity KPIs Performance and related Parameters Objectives After completion of this chapter the participants will be able to: 3 3.1 3.2 3.3 3.4 3.5 3.6 Explain LTE Integrity KPIs performance and related parameters. Describe different LTE integrity KPIs & counters related to Throughput, Latency Explain associated parameters related to different Integrity KPIs Relate QOS and Scheduling profile for integrity optimization Analyze steps to effectible troubleshoot throughput issues Explore related features and associated parameters for throughput enhancement Measure VOLTE effect on throughput and data services KPIs Figure 3-1: Objective of Chapter 3 LZT1381950 R1A © Ericsson AB 2017 - 115 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop Introduction 1 EUTRAN Integrity is a measure of the Network’s ability to provide wireless enduser services with end-user performance quality that meets expectations. This Integrity performance can be divided into the three parts listed below: • EUTRAN Throughput: The speed at which packets can be transferred once the first packet has been scheduled on the air interface. • EUTRAN Latency: The time it takes to schedule the first packet on the air interface, determined from the time it was received in RBS. • EUTRAN Packet Loss: The ratio of lost and transferred packets. This chapter starts with a review on integrity KPIs, their formulas and associated counters will be discussed first, post which, the main focus is intended to be on optimization steps, associated parameters and features, which are important for integrity, mainly throughput KPI will be discussed. 1.1 Average UE Downlink Throughput The average UE downlink throughput KPI formula is illustrated in Figure 3-2 below. The speed at which packets can be transferred once the first packet has been scheduled on the air interface. Average UE DL Throughput [kbps]: = pmPdcpVolDlDrb + pmPdcpVolDlDrbTransUm – S Qci pmPdcpVolDlDrbLastTTIQci pmUeThpTimeDl/1000 Figure 3-2: Review: EUTRAN Throughput KPIs The average UE downlink throughput per QCI for unacknowledged mode RLC and Acknowledged mode RLC KPI is illustrated in Figure 3-3. For cases where ‘pmUeThpTimeDl’ is small and / or where ‘pmPdcpVolDlDrb’ is similar magnitude to ‘pmPdcpVolDlDrbLastTti’ (causing a small numerator from two large values). - 116 - © Ericsson AB 2017 LZT1381950 R1A LTE Integrity KPIs Performance and related Parameters Total DL PDCP UE Throughput shown in Figure 3-3. The speed at which packets can be transferred once the first packet has been scheduled on the air interface. Average UE DL Throughput per Qci (unacknowledged mode RLC) [kbps]: = pmPdcpVolDlDrbTransQci – pmPdcpVolDlDrbLastTTIQci pmDrbThpTimeDlQci/1000 Average UE DL Throughput per Qci (acknowledged mode RLC) [kbps]: = pmPdcpVolDlDrbQci – pmPdcpVolDlDrbLastTTIQci pmDrbThpTimeDlQci/1000 Figure 3-3: Review: EUTRAN Throughput KPIs 1.1.1 Downlink DRB Traffic Volume The counters used to measure the downlink DRB throughput are illustrated in Figure 3-4 below. The last TTI with data is removed since coding may be based on size rather than radio conditions and hence does not impact the end user. Traffic Volume Measured by: pmPdcpVolDlDrbLastTTI Traffic Volume Measured by: pmPdcpVolDlDrbTransUm Traffic Volume Measured by: pmPdcpVolDlDrb DL Transport Time measured by: pmUeThpTimeDl Data arrives to empty DL buffer First data is transmitted to the UE No transmission, buffer not empty (e.g. due to contention) Failed transmission (Block Error) Successful transmission by AM RLC, buffer not empty Transmission by UM RLC Time (ms) Successful transmission by AM RLC, buffer empty Figure 3-4: Review: DL DRB Traffic Volume Measurement LZT1381950 R1A © Ericsson AB 2017 - 117 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop The total DRB traffic volume carried by Acknowledged Mode (AM) RLC that has been transferred and acknowledged by the UE in the downlink direction is measured at the PDCP layer using the ‘pmPdcpVolDlDrb’ ACC counter. The ‘pmPdcpVolDlDrbTransUm’ ACC counter measures the DRB traffic transmitted by the RBS using Unacknowledged Mode (UM) RLC. The downlink DRB traffic volume in the last Transmission Time Interval (TTI) of each packet session is also measured with the ‘pmPdcpVolDlDrbLastTTI’ counter as illustrated in Figure 3-4. This extra measurement is necessary since the coding in the last TTI when the buffer is being emptied may be selected based on size and not the radio conditions, hence the throughput in this last TTI will not impact the end user. The DL Transport Time is measured by the ‘pmUeThpTimeDl’ ACC counter comprises those periods from when the first part of the PDCP SDU of the DL buffer was transmitted on Uu until the buffer is emptied, excluding the TTI emptying the buffer. The details of the downlink DRB throughput counters are given in Figure 3-5 below. Counter Name pmPdcpVolDlDrb pmPdcpVolDlDrbLastTTIQci pmPdcpVolDlDrbTransUm pmUeThpTimeDl Managed Object Description Counter Type EutranCellFDD The total volume (PDCP SDU) on Data Radio Bearers that has been transferred (acknowledged by the UE) in the downlink direction. Continuous measurement for DRBs aggregated to cell level. Unit: kilobit (1 000 bits) ACC EutranCellFDD The total transmitted DL PDCP SDU volume per QCI that are included in the last data transmission in a data burst which is large enough to be transmitted across several TTIs. The bin accumulates the transmitted DL PDCP SDU volume per QCI that are included in the last data transmission in a data burst which is large enough to be transmitted across several TTIs. PDF EutranCellFDD The total volume (PDCP SDU) on Data Radio Bearers for RLC UM that has been transmitted in the downlink direction in the PDCP layer. This is a continuous measurement after Robust Header Compression ACC for DRBs aggregated to cell level Unit: kilobit (1 000 bits) EutranCellFDD The effective DL transport time comprises those periods from when the first part of the PDCP SDU of the DL buffer was transmitted on Uu until the buffer is emptied, excluding the TTI emptying the buffer This is a continuous measurement for UEs aggregated to cell level. Unit: ms ACC Figure 3-5: Review: DL DRB Traffic Volume Counters Note: Also valid for EutranCellTDD - 118 - © Ericsson AB 2017 LZT1381950 R1A LTE Integrity KPIs Performance and related Parameters The counters illustrated in Figure 3-5 are reset at the end of each ROP and collected by the Primary scanner on the RBS. Further counters used by DRB Throughput are described in Figure 3-6. Counter Name Managed Object Description Counter Type pmPdcpVolDlDrbTransQci EutranCellFDD The total volume (PDCP SDU) on Data Radio Bearers that has been transmitted in the downlink direction in the PDCP layer per QCI. Continuous measurement after Robust Header Compression for DRBs aggregated to cell level. PDF pmPdcpVolDlDrbQci EutranCellFDD The total volume (PDCP SDU) that has been transferred (acknowledged by the UE) on Data Radio Bearers in the downlink direction per QCI. The counter accumulates acknowledged DL PDCP volume per QCI. PDF EutranCellFDD The DL transmission time used for DL DRB Throughput per QCI. It comprises of time periods from when the first piece of data in a data burst is transmitted until the second last piece of data in the data burst PDF is transmitted. The bin accumulates the DL transmission time per QCI for each large data burst that needs to be transmitted across multiple TTIs. pmDrbThpTimeDlQci Figure 3-6: Review: DL DRB Traffic Volume Counters 1.1.2 Average UEforUplink Throughput Note: Also valid EutranCellTDD The Average UE Uplink Throughput KPI formula is illustrated in Figure 3-7. The speed at which packets can be transferred once the first packet has been scheduled on the air interface. Average UE UL Throughput [kbps]: = pmUeThpVolUl pmUeThpTimeUl/1000 Average UE UL Throughput per LCG [kbps]: = pmLcgThpVolUlLcg pmLcgThpTimeUlLcg/1000 Figure 3-7: Review: EUTRAN Throughput KPIs LZT1381950 R1A © Ericsson AB 2017 - 119 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop 1.1.3 Uplink DRB Traffic Volume The counters used to measure the uplink DRB throughput are illustrated in Figure 3-8 below. Traffic Volume Measured by: pmUeThpVolUl Contributions from the first 4 and last TTI removed UL Transport Time measured by: pmUeThpTimeUl Data arrives to empty UL buffer Grants sent to the UE Secheduling request sent to RBS Time (ms) First data sent to RBS Send buffer is empty again No receptions, buffer not empty (e.g. due to contention) Failed receptions (Block Error) Successful receptions, buffer not empty Receptions excluded from throughput measurement Successful reception, buffer empty Figure 3-8: Review: UL DRB Traffic Volume Measurement Due to the time difference in the scheduling decision by RBS and the transmission time by the UE the first four receptions are excluded from the DRB traffic volume measurement. The contribution from the last TTI is also removed since the coding when the buffer is being emptied may be selected based on size and not the radio conditions, hence the throughput in this last TTI will not impact the end user. The details of the uplink DRB throughput counters are given in Figure 3-9 below. Counter Name pmUeThpVolUl pmUeThpTimeUl pmLcgThpVolUlLcg pmLcgThpTimeUlLcg Managed Object Description Counter Type EutranCellFDD The UL volume used for UL UE Throughput. It comprises of the MAC SDU volume received on Uu, excluding the volume received in the first 4 data receptions of an UL buffer transfer and the TTI ACC emptying the UL buffer. This is a continuous measurement for UEs aggregated to cell level. Unit: kilobit (1 000 bits) EutranCellFDD The effective UL transport time comprises those periods from when the first part of the PDCP SDU of the UL buffer was received on Uu until the buffer is emptied, excluding the TTI emptying the buffer. This is a continuous measurement for UEs aggregated to cell level. Unit: ms ACC EutranCellFDD The UL volume used for UL DRB Throughput per LCG. It comprises of the MAC SDU volume received on Uu per LCG, excluding the volume received in the first 4 data transmissions and the data transmission emptying the UL buffer. The bin accumulates the UL volume per LCG for large data bursts that needs to be transmitted across multiple TTIs PDF EutranCellFDD The UL transmission time used for UL DRB Throughput per LCG. It comprises of time periods from when the 5th data transmission is received on Uu until the second last data transmission is received. The bin accumulates the UL transmission time per LCG for large data bursts that need to be transmitted across multiple TTIs. PDF Note: Also valid for EutranCellTDD Figure 3-9: Review: UL DRB Traffic Volume Counters - 120 - © Ericsson AB 2017 LZT1381950 R1A LTE Integrity KPIs Performance and related Parameters The counters illustrated in Figure 3-9 are reset at the end of each ROP and collected by the Primary scanner on the RBS. 1.1.4 EUTRAN Latency KPIs The EUTRAN Latency KPIs formulas are illustrated in Figure 3-10 below. The time it takes to schedule the first packet on the air interface, determined from the time it was received in RBS. Downlink Average Latency [ms]: = pmPdcpLatTimeDl pmPdcpLatPktTransDl Downlink Average Latency per QCI [ms]: = pmPdcpLatTimeDlQci pmPdcpLatPktTransDlQci Uplink Latency not measured Figure 3-10: Review: EUTRAN Latency KPIs As illustrated in Figure 3-10 above, there is a KPI formula for overall average downlink latency and one to measure the average latency for each QCI value. The Uplink DRB Latency is not measured in LTE RAN. The ‘pmPdcpLatPktTransDl’ and ‘pmPdcpLatPktTransDlQci’ counters count the number of packets that are used for downlink latency measurement. The latency is measured in msec by the ‘pmPdcpLatTimeDl’ and ‘pmPdcpLatTimeDlQci’ counters as illustrated in Figure 3-11. Packet forwarding during handover may mean that the target cell has to buffer data before it is possible to schedule the UE. For this reason, samples during handover are excluded from the latency counters as this would give a distorted interpretation of the actual latency experienced by the end user, which is not affected by packet forwarding. LZT1381950 R1A © Ericsson AB 2017 - 121 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop The latency Example is illustrated in Figure 3-11 PDCP SDU arrival Samples during Handover excluded No Buffer empty? No receptions, buffer not empty (e.g. due to contention) Successful receptions, buffer not empty Yes Note: Corresponding counter per QCI: pmPdcpLatPktTransDlQci+ t=t0 pmPdcpLatPktTransDl + No Packet sent on Uu? Yes Data arrives to empty DL buffer First data is transmitted to the UE t=t1 t1 – t0 pmPdcpLatTimeDl + aggregated t = (t1 – t0) Note: Corresponding counter per QCI: pmPdcpLatTimeDlQci+ End Figure 3-11: Review: Downlink Latency Measurement The details of the downlink latency counters are given in Figure 3-12 below. Counter Name pmPdcpLatTimeDl pmPdcpLatPktTransDl pmPdcpLatTimeDlQci pmPdcpLatPktTransDlQci Managed Object Description Counter Type EutranCellFDD Aggregated DL Latency for a measurement period. The effective DL Latency time comprises the time from PDCP SDU entering the buffer until the first data has been transmitted to the UE. This measurement for UEs aggregated to cell level and only PDCP SDUs that enter an empty buffer are measured. Unit: msec ACC EutranCellFDD Number of packets for downlink Latency measurements during measurement period. This measurement for UEs aggregated to cell level and only PDCP SDUs that enter an empty buffer are measured. PEG EutranCellFDD Aggregated DL Latency for a measurement period per QCI. The effective DL Latency time comprises the time from PDCP SDU entering the buffer until the first data has been transmitted to the UE. Only latency for PDCP SDUs that enters an empty buffer is measured. Unit: msec PDF EutranCellFDD Number of samples for DL Latency measurements during measurement period per QCI. This counter registers the number of samples for DL Latency measurements during measurement period per QCI. PDF Figure 3-12: Downlink Latency Counters The counters illustrated in Figure 3-12 above are reset at the end of each ROP but only the ‘pmPdcpLatTimeDl’ and ‘pmPdcpLatPktTransDl’ counters are collected by the Primary scanner on the RBS. - 122 - © Ericsson AB 2017 LZT1381950 R1A LTE Integrity KPIs Performance and related Parameters 1.1.5 EUTRAN Packet Loss KPIs The EUTRAN Packet Loss KPI formulas are illustrated in Figure 3-13 below. EUTRAN Packet Loss KPIs The ratio of lost and transferred packets. Downlink Packet Error Loss Rate [%]: = pmPdcpPktDiscDlPelr + pmPdcpPktDiscDlPelrUu + pmPdcpPktDiscDlHo + A pmPdcpPktReceivedDl – pmPdcpPktFwdDl + A Where A = (pmPdcpPktDiscDlEth + pmPdcpPktDiscDlNoUeCtxt) x X100 pmPdcpPktReceivedDl ∑ pmPdcpPktReceivedDl RBS Uplink Packet Loss Rate [%]: pmPdcpPktLostUl = pmPdcpPktLostUl + pmPdcpPktReceivedUl X100 Figure 3-13: EUTRAN Packet Loss KPIs Total number of packets (PDCP SDUs) received in PDCP buffer in the RBS from the SGW is counted by the ‘pmPdcpPktReceivedDl’ counter. The number of these for which no part has been transmitted over the air interface in the downlink direction that are discarded in the PDCP layer due to reasons other than handover is counted with the ‘pmPdcpPktDiscDlPelr’ counter. Those which at least a part has been transmitted over the air interface in the downlink direction but not positively acknowledged, and it was decided that no more transmission attempts will be done, due to reasons other than handover is counted with the ‘pmPdcpPktDiscDlPelrUu’ counter. Those lost due to Handover are counted with the ‘pmPdcpPktDiscDlHo’ counter. LZT1381950 R1A © Ericsson AB 2017 - 123 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop The Downlink Packet Error Loss Rate per QCI and the Uplink Packet Loss Rate are illustrated in Figure 3-14. EUTRAN Packet Loss KPIs The ratio of lost and transferred packets. Downlink Packet Error Loss Rate per Qci [%]: = pmPdcpPktDiscDlPelrQci + pmPdcpPktDiscDlPelrUuQci + pmPdcpPktDiscDlHoQci + B X100 pmPdcpPktDiscDlPelrQci + pmPdcpPktDiscDlPelrUuQci + pmPdcpPktDiscDlHoQci + B + pmPdcpPktTransDlQci B= [pmPdcpPktDiscDlEth + pmPdcpPktDiscDlNoUeCtxt] X pmPdcpPktReceivedDlQci ∑ pmPdcpPktReceivedDlQci RBS Uplink Packet Loss Rate per Qci [%]: pmPdcpPktLostUlQci = pmPdcpPktLostUlQci + pmPdcpPktReceivedUlQci X100 Figure 3-14: EUTRAN Packet Loss KPIs The total number of DRB packets (PDCP SDUs) forwarded to target cell at handover for downlink sessions is counted with the ‘pmPdcpPktFwdDl’. Total number of downlink DRB packets (PDCP SDUs) discarded in the Ethernet part of the RBS are counted with the ‘pmPdcpPktDiscDlEth’ counter. It should be noted that this is a baseband processor resource level counter unlike almost all others that are cell level. - 124 - © Ericsson AB 2017 LZT1381950 R1A LTE Integrity KPIs Performance and related Parameters The counters used to measure downlink packet loss are illustrated in Figure 3-15. Only partially transmitted over air interface due to reasons other than HO: pmPdcpPktDiscDlPelrUu No part transmitted over air interface due to reasons other than HO: pmPdcpPktDiscDlPelr Total packets (PDCP SDUs) received in the PDCP layer in the RBS: pmPdcpPktReceivedDl SGW Lost due to HO: pmPdcpPktDiscDlHo Discarded in Ethernet part of RBS: pmPdcpPktDiscDlEth (NOTE baseband processor level counter) Forwarded to target cell at HO: pmPdcpPktFwdDl Figure 3-15: Downlink Packet Loss Measurement Total packets received in the UL: pmPdcpPktReceivedUl Total packets lost in the UL: pmPdcpPktLostUl Figure 3-16: Uplink Packet Loss Measurement The total number of packets lost in the uplink direction is counted with the ‘pmPdcpPktLostUl’ counter and the total number successfully received packets with the ‘pmPdcpPktReceivedUl’ counter as illustrated in Figure 3-16 above. LZT1381950 R1A © Ericsson AB 2017 - 125 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop The details of the counters used in the Packet Loss KPI formulas are given in Figure 3-17 below. Counter Name Managed Object Description Counter Type pmPdcpPktReceivedDl EutranCellFDD Total number of packets (PDCP SDUs) received in PDCP in the RBS in the downlink direction. ACC pmPdcpPktDiscDlPelr EutranCellFDD Total number of DRB packets (PDCP SDUs) for which no part has been transmitted over the air in the downlink that are discarded due to reasons other than handover or active queue management. ACC pmPdcpPktDiscDlPelrUu EutranCellFDD Total number of packets (PDCP SDUs) for which at least a part has been transmitted over the air in the downlink direction but not positively acknowledged, and it was decided that no more transmission attempts will be done, due to reasons other than handover. Only applicable to DRB packets. ACC pmPdcpPktFwdDl EutranCellFDD Total number of DRB packets (PDCP SDUs) forwarded to target cell at handover for downlink sessions. Packet forwarding is started when the source cell transmits the RRCConnectionReconfiguration message and ends when it receives the end marker or UE Context Release from the target cell. ACC pmPdcpPktDiscDlEth BbProcessingResource Total number of downlink DRB packets (PDCP SDUs) discarded in ACC the Ethernet part of the RBS. pmPdcpPktReceivedUl EutranCellFDD Total number of packets (PDCP SDUs) received in the uplink direction. ACC pmPdcpPktLostUl EutranCellFDD Total number of packets (PDCP SDUs) lost in the uplink direction. ACC Figure 3-17: Packet Loss Counters The counters illustrated in Figure 3-17 above are reset at the end of each ROP and all but the ‘pmPdcpPktFwdDl’ counter is collected by the Primary scanner on the RBS. Counter Name Managed Object Description Counter Type pmPdcpPktDiscDlPelrQci EutranCellFDD Total number of DL packets (PDCP SDUs) per QCI that are discarded in the PDCP layer due to reasons other than handover or active queue management. No part of these packets has been transmitted over the air. PDF pmPdcpPktDiscPelrUuQci EutranCellFDD Total number of DL packets (PDCP SDUs) per QCI for which at least a part has been transmitted over the air but not positively acknowledged due to reasons other than handover, and it was decided that no more transmission attempts will be done. PDF pmPdcpPktDiscDlHoQci EutranCellFDD Total number of DL packets (PDCP SDUs) discarded due to HO per QCI. PDF pmPdcpPktTransDlQci EutranCellFDD Total number of DL packets (PDCP SDUs) transmitted successfully per QCI. PDF PmPdcpPktDiscDlNoUeCtxt BbProcessingResource Number of DRB packets (PDCP SDUs) discarded in downlink due to that the specific Tunnel Endpoint Identifier (TEID) is not mapped to any UE context ACC Figure 3-18: Packet Loss Counters Note: Also valid for EutranCellTDD - 126 - © Ericsson AB 2017 LZT1381950 R1A LTE Integrity KPIs Performance and related Parameters Further counters used in packet loss KPIs are described in Figure 3-18 and Figure 3-19. Counter Name Managed Object Description Counter Type pmPdcpPktReceivedDlQci EutranCellFDD Total number of DL packets (PDCP SDUs) received in PDCP in the RBS per QCI. PDF pmPdcpPktLostUlQci EutranCellFDD Total number of UL packets (PDCP SDUs) lost per QCI. PDF pmPdcpPktReceivedUlQci EutranCellFDD Total number of UL packets (PDCP SDUs) received per QCI. PDF Figure 3-19: Packet Loss Counters 2 Throughput Optimization Steps The LTE end user quality of service is impacted by the cumulative effect of various nodes & interfaces. Figure 3-20 below shows the contributing factors to the end user throughput. • HO Mobility Ping pong • Packet forwarding eNodeB features • MIMO 2x2 4x4 • HARQ eNodeB setup Params Audit • Rec. values Radio Environment • C/I, SINR • Link Adaptation • MCS allocated Peak performance Session Establishment Times Core network / PDN performance • IP delays • Router / #of hops • Packet losses • Session Performance End to End Performance, Core network delays Delay of cell transition BLER at Air I/F: Retransmission rate End server / UE laptop • TCP/IP settings • MTU size E2E Dimensioning • Number of active users/cell • RB allocated • Backhaul TCP/IP issues Available bandwidth per user End user throughput Figure 3-20: Throughput Optimization 2.1 DL Throughput Issues Investigation It is recommended to consider the Accessibility, Mobility and Retainability KPI. Dl Throughput would reduce with increase in number of connected users ’pmRrcConnMax’. Average CQI, % of 64QAM samples and RI indicates DL SINR status. Average CQI should be high (>10%), % of 64QAM sample should be high (>10%), High PRB and PDCCH utilization would impact the DL Throughput ‘pmPrbUtilDl / pmPrbUtilUl’. LZT1381950 R1A © Ericsson AB 2017 - 127 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop High value of DL latency(>9ms) and RLC retransmission (> 1%) would impact DL throughput. High occurrence of ‘pmRadioTbsPwrRestricted’ and ‘pmBadCovEvalReport’ indicates poor DL coverage. DL throughput related issues and investigation is discussed in Figure 3-21 below. Total performance Check all important Accessibly, Mobility and Retainability KPI. RRC connected users Dl Throughput would reduce with increase in number of connected users (pmRrcConnMax). CQI ,64/16QAM,TM modes Average CQI should be high (>10%), % of 64QAM sample should be high (>10%), PRB(DL) and PDCCH utilization High PRB and PDCCH utilization would impact the DL Throughput (pmPrbUtilDl / pmPrbUtilUl). DL Latency and RLC retransmission High value of DL latency(>9ms) and RLC retransmission(> 1%) would impact DL throughput. Power limited UE and No of A2 Search High occurrence of pmRadioTbsPwrRestricted and pmBadCovEvalReport indicates poor DL coverage. Holistic overview of performance makes analysis simpler Av CQI, % 64QAM samples and RI indicates DL SINR status Basic health check up and baseline parameter/feature Audit along with all checks Figure 3-21: Analysis Flow for Dl Throughput Investigation 2.2 Uplink Throughput Investigation To investigate uplink throughput related issues, re-consider the important Accessibly, Mobility and Retainability KPI. Make the baseline Audit for parameters and features. The uplink RSSI may impact the throughput, check the interference counters ‘pmRadioRecInterferencePwrPUCCH’ / ‘pmRadioRecInterferencePwr’. Poor UL SINR conditions would impact UL throughput. Poor UL SINR conditions would impact UL throughput, check the counters ‘pmsinrpucchdistr’ & ‘pmsinrpuschdistr’. High number of power limited UE indicates poor uplink coverage, check the ‘pmRadioTbsPwrRestricted’, ‘pmRadioTbsPwrUNRestricted’ counters. - 128 - © Ericsson AB 2017 LZT1381950 R1A LTE Integrity KPIs Performance and related Parameters The investigation areas for low UL throughput are illustrated in Figure 3-22. Total performance Alarm & Parameter/ Feature Check RSSI Check all important Accessibly, Mobility and Retainability KPI. Holistic overview of performance makes analysis simpler Make Baseline Audit for Parameter and feature High uplink RSSI would impact the throughput (pmRadioRecInterferencePwrPUCCH / pmRadioRecInterferencePwr) % of 16 QAM samples Low usage of 16 QAM modulations scheme in UL would impact the UL throughput PUCCH & PUSCH SINR Poor UL SINR conditions would impact UL throughput Power limited UE (pmsinrpucchdistr & pmsinrpuschdistr ) High number of power limited UE indicates poor uplink coverage (pmRadioTbsPwrRestricted ,pmRadioTbsPwrUNRestricted) Figure 3-22: Analysis Flow for UL throughput Investigation 2.3 Reasons for poor DL Throughput The general optimization strategy to analyze DL throughput issues is described in Figure 3-23 below, it covers the reasons for poor DL throughput. Figure 3-23: Downlink Throughput Optimization LZT1381950 R1A © Ericsson AB 2017 - 129 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop To analyze throughput issues, it should be determined, if the reasons of low throughput in the network is one of the following shown in Figure 3-24 below or a combination of them. › BLER (bad coverage) › Downlink Interference (Bad CQI) › MIMO Parameters › Scheduling algorithm › Low Demand › CQI reporting frequency › Other (VSWR, Backhaul capacity) Figure 3-24: Downlink Throughput Optimization (contd..) 2.4 Reasons for poor UL Throughput The general optimization strategy to analyze the UL throughput issues is described in Figure 3-25 below, it covers reasons for the poor UL throughput. › The general optimization strategy and bad throughput reasons Figure 3-25: Uplink Throughput Optimization - 130 - © Ericsson AB 2017 LZT1381950 R1A LTE Integrity KPIs Performance and related Parameters To analyze the UL throughput issues, it should be determined if the reasons for low throughput in the network is one of the following given in Figure 3-26 below or a combination of more than one issue. › BLER (bad coverage) › Uplink Interference (high RSSI) › Low Power headroom › Scheduling algorithm › Low Demand › Other (VSWR, Backhaul capacity) Figure 3-26: Uplink Throughput Optimization (contd..) 2.5 Throughput Testing Investigation While investigating the throughput issues in the network, the first consideration must be to check the following licenses are applied/enabled/operable and associated parameters are preferably set to optimum values. The licenses for Downlink/Uplink baseband capacity, channel bandwidth in uplink and downlink, 256- QAM, 64-QAM in DL / 16-QAM UL, Dual/Quad Antenna Performance Package for MIMO, Carrier aggregation features should be operable, › The following licenses are applied/enabled/operable. Parameters are set to optimum. › Downlink/Uplink Baseband Capacity › Channel Bandwidth › 256- QAM, 64-QAM DL / 16-QAM UL › Dual/Quad Antenna DL Performance Package › Carrier Aggregation features Figure 3-27: Throughput Testing: Checks LZT1381950 R1A © Ericsson AB 2017 - 131 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop 3 Low Throughput- The Counters to be checked 3.1 PRB Utilization The utilization of downlink Physical Resource Blocks (PRBs) in each eNodeB, relative downlink Physical Resource Block (PRB) pair utilization (total number of used PRB pairs by available PRB pairs) on the Physical Downlink Shared Channel (PDSCH) are collected with the ‘pmPrbUtilDl’ PDF counter, as illustrated in Figure 3-28. The utilization of uplink Physical Resource Blocks (PRBs) in each eNodeB, relative uplink Physical Resource Block (PRB) pair utilization (total number of used PRB pairs by available PRB pairs) on the Physical Uplink Shared Channel (PUSCH) collected with the ‘pmPrbUtilUl’ PDF counter. For Good UL and DL throughput, UL and DL PRB utilization should be low respectively (<80%). › Counters pmPrbUtilUl and pmPrbUtilDl are the PDF counters with 10 ranges each › Concept of weighted average is used to calculate average PRB utilization. Weight of each range (for PDF range[0] from 0 to 10 it has taken as 5) is multiplied by sample in that range and sum of it is divided by total number of samples(in one ROP of 15 minutes it is always going to be 15) Figure 3-28: PRB Utilization - 132 - © Ericsson AB 2017 LZT1381950 R1A LTE Integrity KPIs Performance and related Parameters ‘pmPrbUtilDL’ and ‘pmPRBUtilUL’ is illustrated in Figure 3-29 below.. PRB Utilization in UL: PRB Utilization in Dl : pmPrbUtilUl pmPrbUtilDl A distribution that shows the downlink Physical Resource Blocks (PRB) utilization (total number of used PRB by available PRB) on the Physical Downlink Shared Channel (PDSCH). A distribution that shows the Physical Resource Blocks (PRB) utilization (total number of used PRB by available PRB) on the Physical Uplink Shared Channel (PUSCH). PDF ranges: [0]: 0 % <= Utilization < 10 % [1]: 10 % <= Utilization < 20 % [2]: 20 % <= Utilization < 30 % [3]: 30 % <= Utilization < 40 % [4]: 40 % <= Utilization < 50 % [5]: 50 % <= Utilization < 60 % [6]: 60 % <= Utilization < 70 % [7]: 70 % <= Utilization < 80 % [8]: 80 % <= Utilization < 90 % [9]: 90 % <= Utilization Unit: - PDF ranges: [0]: 0 % <= Utilization < 10 % [1]: 10 % <= Utilization < 20 % [2]: 20 % <= Utilization < 30 % [3]: 30 % <= Utilization < 40 % [4]: 40 % <= Utilization < 50 % [5]: 50 % <= Utilization < 60 % [6]: 60 % <= Utilization < 70 % [7]: 70 % <= Utilization < 80 % [8]: 80 % <= Utilization < 90 % [9]: 90 % <= Utilization Unit: - Condition: One sample should be the utilization during the sample gathering period. Condition: One sample should be the utilization during the sample gathering period. For Good UL and DL throughput- UL and DL PRB utilization should be low respectively ( <80%) Figure 3-29: PRB Utilization 3.2 Ways to reduce PRB utilization There are very limited means to reduce the PRB utilization. However, some of these methods can be tried on the case by case basis, Traffic offloading to less utilized neighboring cells, with IFLB/IFO features. Reduce the control channel resources (Before that check PDCCH utilization). Add Bandwidth (if system is operating < 20Mhz), bandwidth increase would increase number of PRBs, Channel should be equal to licensed bandwidth. Reduce inactivity timer (‘tInactivityTimer’) value so that inactive user can be released early. Some other features like MIMO, CA, 256QAM, 64 QAM, 16QAM, IRC, UL/DL FSS can also be used to reduce PRB utilization, as illustrated in Figure 3-30. › There are very limited means of reducing of PRB utilization .How ever some of these methods can be tried on case by case basis › Traffic offload to less utilized neighboring cells, IFLB, IFO › Reduce control channel resources(Before that check PDCCH utilization) › Add Bandwidth (if system is operating < 20Mhz) as Bandwidth increase would increase number of PRB, Channel should be equal to licensed bandwidth › Reduce inactivity timer (tInactivityTimer) value so that inactive user can be released early › MIMO, CA, 256QAM, 64 QAM, 16QAM, IRC, UL/DL FSS Figure 3-30: Ways to reduce PRB utilization LZT1381950 R1A © Ericsson AB 2017 - 133 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop 3.3 Essential Parameters The essential parameters affecting PRB utilization and throughput need to be optimized are discussed in Figure 3-31 below. › Radio Network MO parameters: EUtranCellFDD › dlChannelBandwidth / ulChannelBandwidth=should be equal to license bandwidth noOfTxAntennas for MIMO= 2 noOfRxAntennas for MIMO= 2 › transmission Mode default is 3, TxDiv and Open Loop Spatial Multiplexing (2x2 MIMO) › pdcchCfiMode (number of OFDM symbols used for PDCCH, 1->3) › maximumTransmissionPower = 460 (for 40 watt site) › partOfSectorPower (for a sector)= 100 › pZeroNominalPucch some UEs need this to be increased or ACK/NACKs are not received successfully on PUCCH. › pZeroNominalPusch some UEs need this to be increased from default or lots of errors seen on PUSCH. – SectorEquipmentFunction › confOutputPower = license power (20/40/60 Watt) Figure 3-31: Essential Parameters 3.4 Other Counters for Throughput Analysis Few more important counters that can be considered for throughput analysis are shown in Figure 3-32 below. › UL Interference on PUCCH / PUSCH: ‘pmRadioRecInterferencePwrPUCCH’ / ‘pmRadioRecInterferencePwr’ › SINR of PUCCH / PUSCH: ‘pmSinrPucchDistr’ / ‘pmSinrPuschDistr’ › RLC ACK/NACK: ‘pmRlcArqUlack’ / ‘pmRlcArqUlNack’ › Transport block PWR restricted: ‘pmRadioTbsPwrRestricted’ / ‘pmRadioTbsPwrUNRestricted’ › Scheduling activity per cell in UL and DL: ‘pmSchedActivityCellDl’ /’pmSchedActivityUeDl’ › Rank distribution MIMO/ SIMO: ‘pmRadioTxRankDistr’ › Number of A2 events(UE in poor coverage) : ‘pmBadCovEvalReport’ Figure 3-32: Counters for throughput analysis - 134 - © Ericsson AB 2017 LZT1381950 R1A LTE Integrity KPIs Performance and related Parameters 3.5 UL Interference To observe UL interference on PUCCH / PUSCH, counters in Figure 3-33. ‘pmRadioRecInterferencePwrPUCCH’ / ‘pmRadioRecInterferencePwr’ should be observed and for SINR of PUCCH / PUSCH counters ‘pmSinrPucchDistr’ / ‘pmSinrPuschDistr’ shown in Figure 3-34 should be observed, for good throughput in UL, UL- RSSI value should be low (< -105 dBm), and few samples should be in poor SINR range. UL Interference on PUCCH* : UL interference on PUSCH : pmRadioRecInterferencePwrPUCCH pmRadioRecInterferencePwr The measured Noise and Interference Power on PUCCH PDF ranges: [0]: N+I <= -121 [1]: -121 < N+I <= -120 [2]: -120 < N+I <= -119 [3]: -119 < N+I <= -118 [4]: -118 < N+I <= -117 [5]: -117 < N+I <= -116 [6]: -116 < N+I <= -115 [7]: -115 < N+I <= -114 [8]: -114< N+I <= -113 [9]: -113 < N+I <= -112 [10]: -112 < N+I <= -108 [11]: -108 < N+I <= -104 [12]: -104 < N+I <= -100 [13]: -100 < N+I <= -96 [14]: -96 < N+I <= -92 [15]: -92 < N+I Unit: dBm/PRB The measured Noise and Interference Power on PUSCH PDF ranges: [0]: N+I <= -121 [1]: -121 < N+I <= -120 [2]: -120 < N+I <= -119 [3]: -119 < N+I <= -118 [4]: -118 < N+I <= -117 [5]: -117 < N+I <= -116 [6]: -116 < N+I <= -115 [7]: -115 < N+I <= -114 [8]: -114< N+I <= -113 [9]: -113 < N+I <= -112 [10]: -112 < N+I <= -108 [11]: -108 < N+I <= -104 [12]: -104 < N+I <= -100 [13]: -100 < N+I <= -96 [14]: -96 < N+I <= -92 [15]: -92 < N+I Unit: dBm/PRB Figure 3-33: RSSI SINR of PUCCH: SINR for PUSCH: pmSinrPucchDistr pmSinrPuschDistr Distribution of the SINR values calculated for PUCCH . Distribution of the SINR values calculated for PUSCH . › PDF ranges: › [0]: SINR <= -15 › [1]: -15 < SINR <= -12 › [2]: -12 < SINR <= -9 › [3]: -9 < SINR <= -6 › [4]: -6 < SINR <= -3 › [5]: -3 < SINR <= 0 › [6]: 0 < SINR <= 3 › [7]: 3 < SINR › Unit: dB Condition: Each SINR value for PUCCH per UE calculated on a TTI basis yields one sample in the distribution. › PDF ranges: › [0]: SINR <= -5 › [1]: -5 < SINR <= -2 › [2]: -2 < SINR <= 2 › [3]: 2 < SINR <= 6 › [4]: 6 < SINR <= 10 › [5]: 10 < SINR <= 14 › [6]: 14 < SINR <= 17 › [7]: 17 < SINR › Unit: dB Condition: Each SINR value for PUSCH per UE calculated on a TTI basis yields one sample in the distribution. For good throughput in UL - FEW Samples in poor SINR ranges Figure 3-34: UL SINR LZT1381950 R1A © Ericsson AB 2017 - 135 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop 3.6 UL RLC NACK The total number of successful RLC PDU transmissions (ACKs) in the uplink direction is observed with ‘pmRlcArqUlack’. The counter does continuous measurement for Radio Bearers aggregated to cell level. The total number of unsuccessful RLC PDU and RLC PDU segment transmissions (NACKs) in the uplink direction can be observed with counter ‘pmRlcArqUlNack’. For good throughput in UL, UL RLC NACK ratio should be low. And below Figure 3-35 illustrated the counters. RLC ACK: pmRlcArqUlack RLC NACK: pmRlcArqUlNack › The total number of successful RLC PDU transmissions (ACKs) in the uplink direction. › Condition: Continuous measurement for Radio Bearers aggregated to cell level.. › The total number of unsuccessful RLC PDU and RLC PDU segment transmissions (NACKs) in the uplink direction. › Condition: Continuous measurement for Radio Bearers aggregated to cell level. For good throughput in UL –UL RLC NACK RATIO should be low Figure 3-35: UL RLC NACK 3.7 Power restricted transport block in UL The number of Transport Blocks on MAC level scheduled in uplink where the UE was considered to be power limited. A Transport Block is considered to be power limited when the estimated required transmit power is higher than the UE maximum transmit power, counter is ‘pmRadioTbsPwrRestricted’. The scanner not included in any predefined scanner. ‘pmRadioTbsPwrUNRestricted’ are the number of Transport Blocks on MAC level scheduled in uplink where the UE was NOT considered to be power limited. A Transport Block is considered to be power limited when the estimated required transmit power is higher than the UE maximum transmit power. Figure 3-36 shows these counters. - 136 - © Ericsson AB 2017 LZT1381950 R1A LTE Integrity KPIs Performance and related Parameters Ratio of power limited blocks to non-power limited blocks should be low for good throughput in UL, ratio should be < 50%. Transport block PWR restricted: pmRadioTbsPwrRestricted › The number of Transport Blocks on MAC level scheduled in uplink where the UE was considered to be power limited. › Unit: › Condition: A Transport Block is considered to be power limited when the estimated required transmit power is higher than the UE maximum transmit power. › Counter type: PEG › Scanner: Not included in any predefined scanner Transport block PWR restricted: pmRadioTbsPwrUNRestricted › The number of Transport Blocks on MAC level scheduled in uplink where the UE was NOT considered to be power limited. › Unit: › Condition: A Transport Block is considered to be power limited when the estimated required transmit power is higher than the UE maximum transmit power. › Counter type: PEG › Scanner: Not included in any predefined scanner Ratio of power limited to non power limited blocks should be low For Good throughput in UL- Ratio should be low(< 50%) Figure 3-36: Power restricted transport block-UL 3.8 CQI Range As shown in Figure 3-37 the counter ‘pmRadioUeRepCqiDistr’ gives CQI distribution in DL and the counter ‘pmRadioUeRepCqiDistr; is a PDF counter and it gives value range from 0 to 15. For good throughput in DL- average CQI should be high (>10). CQI is a feedback mechanism from UE to eNB for downlink › pmRadioUeRepCqiDistr gives CQI distribution in DL › pmRadioUeRepCqiDistr is a PDF counter and it gives value in range from 0 to 15 › CQI 1-6 map to QPSK › CQI 7-9 map to 16QAM › CQI 10-15 map to 64QAM For good throughput in DL- Average CQI should be high (>10) Figure 3-37: CQI LZT1381950 R1A © Ericsson AB 2017 - 137 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop 3.9 MIMO RANK distribution usage Rank distribution MIMO counter is ‘pmRadioTxRankDistr’, discussed in Figure 3-38. The transmission mode / rank distributions give more detailed information on how much each transmission mode and rank is used. For good throughput in DL high samples of MIMO Rank 2 should be observed. Rank distribution MIMO/ SIMO : pmRadioTxRankDistr The transmission mode / rank distributions gives more detailed information on how much each transmission mode and rank is used. PDF ranges: [0]: Transmit diversity same information both branches [1]: Open Loop SM Rank 1 SIMO [2]: Open Loop SM Rank 2 MIMO [3]: Closed Loop SM rank 1 [4]: Closed Loop SM rank 2 Unit: Counter type: PDF Scanner: Not included in any predefined scanner SM stands for Spatial Multiplexing For Good throughput in DL- High samples of MIMO Rank 2 Figure 3-38: MIMO RANK distribution usage 3.10 Modulation Scheme usage DL Different modulation scheme samples can be observed with the counters discussed in Figure 3-39 for DL and in Figure 3-40 for uplink. The percentage of QPSK Samples, 16 QAM Samples, 64QAM Samples & 256QAM samples (if HOM implemented) can be calculated and for good DL throughput, high usage of 256 QAM, 64 QAM and 16 QAM scheme in DL should be observed. › QPSK Samples (%) = 100 * (pmMacHarqDlAckQpsk/ (pmMacHarqDlAckQpsk + pmMacHarqDlAck16qam + pmMacHarqDlAck64qam)) › 16 QAM Samples (%) = 100 * (pmMacHarqDlAck16qam / (pmMacHarqDlAckQpsk + pmMacHarqDlAck16qam + pmMacHarqDlAck64qam)) › 64QAM Samples(%) = 100 * (pmMacHarqDlAck64qam/ (pmMacHarqDlAckQpsk + pmMacHarqDlAck16qam + pmMacHarqDlAck64qam)) For good DL throughput- High usage of 256 QAM, 64 QAM and 16 QAM scheme in DL Figure 3-39: Modulation scheme usage DL - 138 - © Ericsson AB 2017 LZT1381950 R1A LTE Integrity KPIs Performance and related Parameters Modulation scheme used in UL is illustrated in Figure 3-40 QPSK Samples (%) = 100 * ((pmMacHarqUlSuccQpsk/ (pmMacHarqUlSuccQpsk+ PmMacHarqUlSucc16qam )) 16 QAM Samples (%) =100 * ((pmMacHarqUlSucc16qam/ (pmMacHarqUlSuccQpsk+ PmMacHarqUlSucc16qam )) For good UL throughput - High usage of 64/16 QAM scheme in UL Figure 3-40: Modulation scheme usage UL 4 Throughput Optimization- UE essential checks UE Subscriber profile also play an important role in high throughput, subscribed charging characteristics, Aggregate Maximum Bit Rate (AMBR), Max requested bandwidth in Downlink, Max requested bandwidth in Uplink, RAT frequency selection priority, APN configuration profile, default Context Identifier (default APN for the EPS User), APN Configuration (every APN associated to the EPS Users), QOS parameter setting, transport Network characteristics. These profiles are important for throughput optimization, Figure 3-41. UE Subscriber profile (End User (EPS User) subscription data is stored in the HSS) › Subscribed Charging Characteristics › Aggregate Maximum Bit Rate (AMBR) Max requested bandwidth in Downlink Max requested bandwidth in Uplink › RAT frequency selection priority › APN configuration profile: Default Context Identifier (default APN for the EPS User) APN Configuration (every APN associated to the EPS User) › QOS Parameter setting › Transport Network issue Figure 3-41: Throughput Optimization-essential checks LZT1381950 R1A © Ericsson AB 2017 - 139 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop 4.1 QOS Framework The LTE Quality of Service (QoS) Handling coordinates and assigns the appropriate QoS to other functions in LTE RAN. The RBS maps QCIs (Quality of Service Class Identifiers) to priorities for different Data Radio Bearers (DRBs) in the LTE radio interface and different data flows in the transport network. The LTE QoS Handling complies with the 3GPP QoS concept. It provides service differentiation per UE by support of multiple parallel bearers. To provide service differentiation in the uplink, traffic separation must be ensured between the different data flows within the UE. This is done by offering an operatorconfigurable mapping between QCIs and LCGs (Logical Channel Groups, also sometimes referred to as radio bearer groups). Signaling Radio Bearers (SRBs) are assigned higher priority than Data Radio Bearers (DRBs). SRB1 has higher priority than SRB2. For the UL, the transport network benefits from QoS by mapping QCI to DiffServ Code Point (DSCP) in the RBS. This enables the transport network to prioritize between its different data flows over the S1 interface in the uplink and over the X2 interface for the downlink data in case of Packet Forwarding. For the DL, a similar mapping is performed in the S-GW for the S1 DL data. If a user has multiple bearers with different QCI, these users will be separated in the radio network into different bearers. This separation will achieve benefits for the end-user QoS as it removes the risk that one service such as file download would block the traffic for a voice call. QoS Handling is based on mapping QCIs received from the core network to RBS-specific parameters. This makes it possible to have different priorities and DSCP values. - 140 - © Ericsson AB 2017 LZT1381950 R1A LTE Integrity KPIs Performance and related Parameters The LTE QoS Handling is realized by a central function in the RBS, which directly influences the radio and transport network performance. The QOS process is illustrated in Figure 3-42 below. DSCP: DiffServ Code Point OSS: Operations Support System QoS: Quality of Service QCI: QoS Class Indentifier RC: Radio and Core Core Network - absPrioOverride - priority resourceType logicalChannelGroupRef QCI Table, Example of population Standardized QCIs OSS-RC QoS parameters QCI Table QCI RT Prio LCG DSCP 1 GBR 2 2 46 2 GBR 4 1 36 : : : : : 9 Non-GBR 9 3 12 10 3 0 10-255 parameters QCI, ARP paArpOverride Scheduler QCI table •QoS configuration QoS translation Priorities LCGs QoS Handling DL Packet Forwarding (X2) UL/DL (Radio Interface) Radio Network UL (S1) Transport Network UE Figure 3-42: QoS Framework In uplink, the distribution of the granted resources is done per logical channel internally within the UE using the rate control function. The RBS maps the QCI to LCG and informs the UE about the association of a logical channel to a LCG and the logical channel priority for each logical channel. The reason for using LCGs is that the Buffer Status Report (BSR) is sent per LCG and not per logical channel. This reduces the uplink signaling load. Standardized QCIs (1-9) are used, according to 3GPP TS 23.203. Non-standardized QCIs (10-256) are all given the same priority, which shall be lower compared to priorities for the standardized QCIs. For the uplink, the priorities are translated to logical channel priorities and sent to the UE, which may differentiate/prioritize between its logical channels. Mapping QCIs to Logical Channel Groups (LCGs) can be configured in OSS-RC and enables traffic separation in the uplink. There are three LCGs (1-3) available. By default, LCG 1 is assigned to all QCIs. The UE performs prioritization between its own logical channels. This is referred to as UE rate control. LZT1381950 R1A © Ericsson AB 2017 - 141 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop The QOS based scheduling example is discussed in Figure 3-43. Prio LCG Abs Prio Override QCI 1 2 1 HI_PRIO_OVERRIDE QCI 2 3 1 NO_OVERRIDE QCI 3 4 2 NO_OVERRIDE Scheduled before all other QCIs* Scheduled according to scheduling algorithms ** RF, PF UE Rate Control *In UL, the established bearer with highest prio represents the abs prio for the LCG **In UL, the established bearer with highest prio represents the scheduling algorithm for the LCG Figure 3-43: QoS Aware Scheduler – Absolute priority example 4.2 Scheduling Algorithm Scheduler Configuration The two main scheduling algorithms are ‘Resource Fair’ and ‘Minimum Rate Proportional Fair’. The later one works together with the QoS Aware Scheduler feature. This scheduling algorithm is based on channel quality information and scheduled data rates. It allows better sharing of RAN, RBS and baseband resources between different radio bearers, thus providing a trade-off between user demands and system performance. Considering the fact, in a mobile network environment, different users will experience different radio conditions at a certain time; this scheduling algorithm prioritizes users experiencing good radio quality, thus leading to a higher throughput when compared to the basic scheduling method (Resource Fair). However, to avoid some users being allocated too few or no resources due to poor channel quality, fairness is provided also by taking into account the average rate in scheduler prioritization. The scheduler needs as an input the channel quality of the users in the cell and the average rates of their flows. Therefore, it monitors the Channel State Information (CSI) from the UE and measurements performed in the RBS. The average rate is calculated based on previous transmissions for the flows. When resources are divided among users the prioritization is based on a weighted measure of the channel quality and average rate. - 142 - © Ericsson AB 2017 LZT1381950 R1A LTE Integrity KPIs Performance and related Parameters In many scenarios, cell capacity can be increased by using the low fairness versions of the proportional fair scheduling algorithms. These algorithms increase the share of the resources given to users with good channel conditions. This leads to an overall increase in capacity. Different scheduling algorithm are illustrated in Figure 3-44. › Proportional Fair Scheduling algorithms parameter: QciProfilePredefined.schedulingAlgorithm › Equal Rate › Proportional Fair High › Proportional Fair Medium › Proportional Fair Low › Max C/I › Min rate in UL/DL. Parameters: – QciProfilePredefined.ulMinBitRate – QciProfilePredefined.dlMinBitRate Figure 3-44: Scheduler Configuration The feature also helps in giving a consistent cell edge throughput with the minimum rate Quality of Service (QoS) characteristic. The basic concept is that flows experiencing a bit rate lower than their configured minimum bit rate will be prioritized before flows that have a bit rate above their configured minimum rate. As mentioned above, Proportional Fair scheduling feature takes into account both scheduled data rate and radio channel quality. The tradeoff between user fairness and the system performance can be tuned by Channel Quality Fraction (CQF). The CQF controls how big portion of channel quality should contribute to a user's priority. With increasing fairness, the scheduler will try to bring all the users to the same received data rate range. On the other hand, system performance could degrade, as more resources need to be allocated to users in bad channel conditions to give them as high data rate as users in good channel conditions. In the most unfair setting (Max C/I), then resources are spent on good channel users yielding a high system throughput. However, cell edge users are disfavored. To control the CQF’s impact on the scheduler, five different scheduling algorithms can be configured, see Figure 3-44 above. These 5 different types of configuration profiles will allow the operator to set the network behavior to a scenario that varies from providing higher fairness to the users (equal rate) to a higher capacity (Max C/I), which provides a higher throughput. LZT1381950 R1A © Ericsson AB 2017 - 143 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop This is detailed in the charts of the below Figure 3-45. Opportunistic scheduling enables increased capacity and increased cell edge throughput in higher load scenarios User Throughput Higher fairness Higher capacity Average Cell Throughput Higher fairness Higher capacity Figure 3-45: Proportional Fair Scheduling Benefits Feature affecting UL/DL Throughput 5 Important LTE features affecting UL/DL throughput is shown in Figure 3-46 below. Figure 3-46: LTE Features enhancing UL/DL Throughput - 144 - © Ericsson AB 2017 LZT1381950 R1A LTE Integrity KPIs Performance and related Parameters 5.1 256 QAM DL The feature provides support of 256 QAM in DL. This allows much higher data rate for 256-QAM capable UEs when their DL link quality is very good. Compared to 64 QAM, this feature increases the DL user peak rate (MAC layer) from 150 Mbps to about 195 Mbps with 20 MHz bandwidth and 2-layer MIMO. With 3-carrier carrier aggregation, the DL user peak rate can reach about 587 Mbps. In L16A, 256 QAM is supported in transmission mode 1, 2, 3 and 4, while the highest rank is limited to 2. To support the feature, UE must support 256 QAM as specified in 3GPP Release 12: This includes reporting 256 QAM capability, can be configured to use CQI table for 256 QAM and can demodulate and decode transport blocks using 256 QAM. 256 QAM is supported only with rank of 1 and 2. If 4x4 MIMO license is enabled/activated and a UE is capable of 4x4 MIMO, 256 QAM is not supported for the UE. In case of cross-DU CA, the UE peak rate of 587 Mbps can’t be supported if the CA capable UE has PCell on one DU and two SCells on the other DU. The license MO instance name is ‘Dl256Qam’. To activate the feature, one needs to enable and activate the feature license. The attribute ‘dl256QamEnabled’ should be set to TRUE. This attribute is used to enable or disable the feature on cell level. Since low transmit EVM (error vector magnitude) is needed to achieve good 256 QAM performance, the attribute ‘radioTransmitPerformanceMode’ is used to configure the desired radio performance mode. This attribute sets the desired mode, not the actual mode. Whether the radio can operate at the desired mode also depends on the configured carrier powers for all carriers transmitted from the radio. If supported, radios can be configured to work in one of the four radio transmit performance modes: DEFAULT: Legacy, radio shall behave the same as in L15B. Performance for 64 QAM may not be optimal for some radios. ‘QAM_64_BOOST’: The requested power headroom is for 64-QAM performance. ‘QAM_256_BOOST’: The requested power headroom is for 256-QAM performance, which is more than that for 64-QAM. LZT1381950 R1A © Ericsson AB 2017 - 145 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop MAX_POWER_CLASS: The requested power headroom is the maximum the radio can provide. This could provide good 256-QAM performance, but the radio will not be power efficient. In addition, too much headroom will cause performance issues as the radio transmit signal will not be operating within its optimal dynamic range. The attribute ‘dl256QamStatus’ indicates whether radio has enough power headroom for 256 QAM. If ‘dl256QamStatus’ = SUPPORTED, it means radio has enough power headroom for 256 QAM. If ‘dl256QamStatus’ = DEGRADED, it means 256 QAM can be supported, but the radio power headroom may not be enough to achieve 256 QAM peak rate. If ‘dl256QamStatus’ = NOT_SUPPORTED, it means radio doesn’t have enough power headroom to support 256 QAM, or the 256 QAM license is not enabled, or the ‘dl256QamEnabled’ is not set to TRUE. As describer earlier, the lack of radio power headroom can be due to carrier power configuration or other factors like temperature. In order to achieve 256 QAM peak rate, please ensure that the total configured carrier power is less than the max radio transmit power and the value of ‘dl256QamStatus’ is SUPPORTED. ManagedElement +-ENodeBFunction +-SectorCarrier radioTransmitPerformanceMode (DEFAULT, QAM_64_BOOST, QAM_256_BOOST, MAX_POWER_CLASS) A desired radio transmit operating mode, which is tailored to achieve good throughput performance for a given highest modulation in downlink. +-EUtranCellFDD / EUtranCellTDD dl256QamEnabled {True, False} Enable or disable 256 QAM support in DL dl256QamStatus (Read Only) {NOT_SUPPORTED, DEGRADED, SUPPORTED} Indicate whether 256 QAM can be supported and whether radio has enough power headroom Figure 3-47: 256 QAM DL Parameters The target EVM for 256-QAM is 2%. The requested power headrooms are stored in radio data base, and are used to set the radio according to the configured radio performance mode. In 16A, the power headroom stored in radio data base may not map to the target EVM for some radios or for some configurations (like narrow bandwidth). - 146 - © Ericsson AB 2017 LZT1381950 R1A LTE Integrity KPIs Performance and related Parameters Although the radio performance mode is per sector-carrier, the support of 256 QAM is on the cell level. In case of combined cell with multiple radios, if any radio can’t support radio performance mode, 256 QAM is not supported by the cell. Similarly, if all radios support radio performance mode but one radio doesn’t have enough power headroom for 256 QAM, then 256 QAM is not supported by the cell. The feature can be supported in mixed mode where one radio transmits and receives signals for LTE and other technologies. However, LTE throughput degradation may be observed in some cases. The attribute ‘radioTransmitPerformanceMode’ is per sector-carrier. If multiple sector-carriers are supported by a single radio, the same ‘radioTransmitPerformanceMode’ should be configured. If different radio performance modes are configured for different sector-carriers in this case, the highest mode is used by the radio (the order of the modes from high to low is: MAX_POWER_CLASS, QAM_256_BOOST, QAM_64_BOOST, and DEFAULT). For example, sector-carrier1 and sector-carrier2 are supported by the same radio, while the ‘radioTransmitPerformanceMode’ is configured as “DEFAULT” for sector-carrier1 and ‘QAM_256_BOOST’ for sector-carrier2. In this case, the radio will use ‘QAM_256_BOOST’ mode. However, the performance mode displayed on the user interface for sector-carrier1 is still “DEFAULT”. The feature supports two modulation modes: the legacy mode in which the highest modulation is 64 QAM, and the higher order modulation (HOM) mode in which the highest modulation is 256 QAM. In HOM mode, 256 QAM can be selected whenever link quality permits. The dynamic mode switch between the two modes is supported. When a 256 QAM capable UE is in legacy mode, it can use any transmission mode that is supported in our product so far. For a 256 QAM capable UE, if it is in legacy mode and is in any of the four Tx modes listed above, it can be switched to HOM mode when the DL link quality is good enough. If it is in legacy mode and is in a Tx mode other than the four modes listed above, switching to HOM mode is not allowed. If any UE is in HOM mode and is in a Tx mode other than the four modes listed above, the UE will be put into legacy mode. If 4x4 MIMO license is enabled and a UE is 4x4 MIMO capable, the UE will not be put in HOM mode. This is mainly due to development cost and uncertainty about UE support. If 4x4 MIMO license is disabled and a UE is capable of 4x4 MIMO and 256 QAM, we will limit the UE’s reported rank to 2. LZT1381950 R1A © Ericsson AB 2017 - 147 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop The tables shown in Figure 3-48 for MCS index to transport block size index is defined to be used in HOM mode. Here we see that the modulation order steps up to 8 bits per symbol for the higher MCS indexes. MCS Index Modulation Order TBS Index I MCS Qm I TBS 0 2 0 1 1 2 2 2 2 2 4 2 3 3 2 6 4 2 4 4 2 8 5 2 5 5 4 10 6 2 6 6 4 11 7 2 7 7 4 12 8 2 8 8 4 13 9 2 9 9 4 14 10 4 9 10 4 15 11 4 10 11 6 16 12 4 11 12 6 17 13 4 12 13 6 18 14 4 13 14 6 19 15 4 14 15 6 20 16 4 15 16 6 21 17 6 15 17 6 22 18 6 16 18 6 23 19 6 17 19 6 24 18 20 8 8 25 MCS Index I MCS Modulation Order TBS Index Qm I TBS 0 2 0 1 2 2 2 3 20 6 21 6 19 21 22 6 20 22 8 28 23 6 21 23 8 29 24 6 22 24 8 30 25 6 23 25 8 31 26 6 24 26 8 32 8 33 27 6 25 27 28 6 26 28 2 29 4 30 6 31 8 29 2 30 4 31 6 reserv ed 27 reserv ed Figure 3-48: Link Adaptation Tables 5.2 64 QAM Uplink 64-QAM Uplink is an optional feature that enables higher order modulation on the PUSCH. Enabling the feature increases the number of bits per symbol used in LTE physical layer resources and thus improves throughput. Support of 64QAM in PUSCH by Category 5 UEs is defined in 3GPP since Release 8. Support of 64QAM in PUSCH by UL Category 5 and 13 UEs is defined in 3GPP Release 12. UL MU-MIMO, UL Carrier Aggregation and Clustered PUSCH will work together with 64-QAM Uplink. There should be no limitation on the number of 64QAM users that this feature should support. MCS with 64QAM modulation will only be scheduled for UEs that support it. Furthermore, a power offset to ‘p0_NominalPusch’ can be used for 64QAM capable UEs in order to have a higher target SINR for selecting the most efficient MCS. - 148 - © Ericsson AB 2017 LZT1381950 R1A LTE Integrity KPIs Performance and related Parameters 64QAM is selected based on estimated uplink channel quality. Since UE is power limited, 64QAM may be applied in good radio conditions with the knowledge of UE’s power headroom. The overall uplink cell throughput and spectral efficiency is expected to increase, however the amount of the increase depends on the ratio of 64QAM users and their channel characteristics. Ideally it allows for single user uplink peak rate and spectral efficiency as below: • ~75 Mbps in 20 MHz FDD, while ~55 Mbps in case of 16QAM • 3.75 bits/s/Hz, while 2.75 bits/s/Hz in case of 16QAM – More information transferred in each UL radio symbol (total 6 bit) › 50% more compared to 16-QAM – Higher peak bit rate in good radio conditions 64-QAM › Operator benefit – Improved cell capacity – Improved peak rates 6 bits/symbol 50% higher UL peak rates and increased cell capacity Figure 3-49: 64-QAM Uplink (FDD/TDD) › 64QAM modulation increases the number of bits that can be transmitted per symbol. › Allows uplink peak rate of up to 70 Mbps* for 64 QAM capable UEs in FDD 20 MHz bandwidth. › Allows uplink peak rate of up to 13.77 Mbps* for 64QAM capable UEs in TDD 20 MHz bandwidth with UL/DL configuration 2. › Increases uplink throughput for 64 QAM capable UEs in good coverage areas. Bits per symbol vs modulation *Note : This is based on that the configuration is 96 SBs in TTI without PRACH, and 90 SBs in TTI with PRACH, with SRS disabled Number of bits per symbol 7 6 5 4 3 2 1 0 QPSK 16QAM 64QAM Modulation Figure 3-50: 64 QAM UL Benefits LZT1381950 R1A © Ericsson AB 2017 - 149 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop 5.3 64 QAM UL Parameters New MOM attribute ‘puschPwrOffset64qam’ is introduced considering that the majority of the UEs are not capable of 64 QAM in Uplink. With this in mind it is desired to keep ‘pZeroNominalPusch’ unchanged. The ‘puschPwrOffset64qam’ is used to control the received SINR target and enable maximum throughput for 64QAM, which in some configurations is not possible otherwise. The ‘puschPwrOffset64qam’ is an offset to ‘pZeroNominalPusch’ in close-loop power control. For 64QAM capable UEs (regardless of the actual modulation), the PUSCH Tx power per resource element is calculated by (‘pZeroNominalPusch’ + ‘puschPwrOffset64qam’). When setting ‘puschPwrOffset64qam’, the following need to be considered: setting of ‘pZeroNominalPusch’, number of RX antennas, and the desired performance for 64QAM capable UEs. For 2 Rx, ‘pZeroNominalPusch’ + ‘puschPwrOffset64qam’ needs to be -100 dBm or higher to achieve maximum throughput. For more than 2 Rx, this needs to be adjusted considering the diversity gain, in ideal case 3 dB and in reality the gain should be less than 3 dB. If UL MU-MIMO is enabled in the eNB, ‘puschPwrOffset64qam’ should be set to 0 to avoid extra interference when pairing UEs w/wo 64QAM capability. When this feature is enabled and the ‘puschPwrOffset64qam’ > 0 dB, 64QAM capable UEs will generate more interference within the cell and towards neighboring cells. The throughput of UEs without 64QAM capability will be impacted. This behavior is expected due to two reasons: • Higher PUSCH power for 64QAM users will generate more interference to the other users. • The higher PUSCH power also gives better SINR thus 64QAM users have advantage in frequency selective scheduling. The attribute ‘ul64qamEnabled’ should be set to TRUE. This attribute is used to enable or disable the feature on cell level. - 150 - © Ericsson AB 2017 LZT1381950 R1A LTE Integrity KPIs Performance and related Parameters New counters and performance indicators are introduced by this feature in order to monitor the performance of 64QAM in PUSCH. It also changes the calculation of existing performance indicators. +- EUtranCellFDD/EUtranCellTDD puschPwrOffset64Qam {0..8} Unit: 1 dB def.: 0 dB Power offset to the target received PUSCH Power Spectral Density (PSD) for 64-QAM-capable UEs. Used to control the received SINR target. It is used to enable maximum throughput for 64QAM, which in some configurations is not possible. For 2Rx, pZeroNominalPusch + puschPwrOffset64qam needs to be 100 dBm or higher to achieve maximum throughput. +-EUtranCellFDD/EUtranCellTDD ul64qamEnabled {True, False} def.: True Enable or disable 64 QAM support in UL Figure 3-51: 64 QAM UL Parameters 5.4 Ericsson Lean Carrier The cell specific reference signals (CRS) are UE known symbols that are inserted within the OFDM time and frequency grid. The reference symbols are used by the UE for intra and inter frequency mobility measurements, downlink channel estimation, Automatic Gain Control and Automatic Frequency Control. In 3GPP, the Cell specific Reference Signals (CRSs) are specified to always be transmitted. The CRSs contribute to severe inter-cell interference towards neighboring cells. Specifically, in low load conditions, the CRS induced interference is often the dominating factor that limits DL throughput performance. The Ericsson Lean Carrier feature addresses this problem by only sending CRSs when and where they are needed to support the traffic, while minimizing the harmful interference. Transmissions of reference signals are limited to a bare minimum without negative impact of network KPI e.g. mobility and call setup success rate. The overall reduction of CRS transmissions depends on cell load. Up to 80% reduction in emitted reference signal power can be achieved. LTE devices requires reference signals to function. Today, excessive LTE reference signaling causes interference between cells limiting network performance. Continuous reference signal transmission is done across all subframes and PRBs at all times. LZT1381950 R1A © Ericsson AB 2017 - 151 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop In contrast, Ericsson Lean Carrier adapts 5G lean concepts to reduce LTE reference signaling and the associated inter cell interference, improving the mobile experience. Ericsson Lean Carrier Transmits Reference Signals in subframes and in PRBs only when required. Ericsson Lean Carrier: Ericsson Lean Carrier: Current LTE reference signaling User Data causes interference between cells Throughput Ericsson Lean Carrier dynamically reduces reference signaling by up to 80%, reducing inter-cell interference Interference Interference USER DATA USER DATA User Data Throughput Figure 3-52: Ericsson Lean Carrier Applying 5G concepts to today’s 4G LTE Reference signals are muted all times except for the following: • Reference signals are always transmitted on the 6 center PRB. This ensures that devices can evaluate the cell for cell selection purposes. • Reference signals are always transmitted over the full bandwidth at the following occasions: - SIB sub-frames - Paging sub-frames - During connection set up until DRX configuration is sent to UE. All devices are configured with DRX. - When a UE is scheduled - During “on-duration” in DRX cycle, since this is the time when the UE is monitoring the PDCCH. - Configurable to enable more sub-frames with full reference signal bandwidth pre and post “on-duration” With the Ericsson Lean Carrier function, the CRS are muted in subframes where they are not expected by the UEs. This reduces the overall inter cell interference and improves downlink user bit rates. It should be noted that CRS are not muted in the six center sub-carriers. - 152 - © Ericsson AB 2017 LZT1381950 R1A LTE Integrity KPIs Performance and related Parameters 5.5 Carrier Aggregation The 'Carrier Aggregation' Value Package contains 7 functions, Figure 3-53. The 'DL 4CC Carrier Aggregation' function is new in L17.Q1. The optional feature 4CC DL Carrier Aggregation Extension enables carrier aggregation of up to four DL component carriers. With Carrier Aggregation a UE may use resources from multiple cells. With 4CC DL Carrier Aggregation Extension, the UE may use up to 4 DL CCs and 1 UL CC at the same time. The main benefits of 4CC DL Carrier Aggregation Extension are as follows: • Enables bandwidth aggregation of three FDD carriers in DL, up to 60 MHz aggregated BW. • More efficient use of scattered spectrum. • Further increased UE downlink bitrate. For VoLTE Interaction: • If Enhanced PDCCH Link Adaption (PDCCH LA) is enabled and a VoLTE bearer is configured, no data (VoLTE or other) will be scheduled on SCells. This may give a throughput degradation for CA capable UEs during VoLTE calls • If Enhanced PDCCH LA is not enabled, there is no special handling of VoLTE, and VoLTE traffic can be scheduled either on the PCell or an activated SCell. A UE configured with a SCell consumes more resources, one for the PCell and one for each configured SCell, than a UE that only has a PCell, yet it is still counted as one connected UE from the connected UE license perspective. As the 4CC DL Carrier Aggregation Extension feature allows up to 3 configured SCells per UE, the loading limit can be reached by fewer UEs than was previously possible. Existing mechanisms with the right configuration ensure that SCell configuration is stopped before the loading limit is reached and therefore prevents carrier aggregation UEs from blocking the attachment of new UEs to the system. The PUCCH resource dimensioning must be considered. UE may use up to 4 downlink + 1 uplink CA supported only for FDD at the same time. As result of the feature implementation further increased UE peak rates are expected. However, the benefit decreases as the number of users with traffic activity increases. LZT1381950 R1A © Ericsson AB 2017 - 153 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop The UL UE peak throughput is reduced in good radio conditions because of extra HARQ ACK/NACK for the downlink transmission for the SCells to be sent through PUSCH when the UE is configured with multiple SCells. Ericsson Lean Carrier: Ericsson Lean Carrier: Current LTE reference signaling User Data causes interference between cells Throughput Ericsson Lean Carrier dynamically reduces reference signaling by up to 80%, reducing inter-cell interference Interference Interference USER DATA USER DATA User Data Throughput Figure 3-53: Ericsson Lean Carrier Applying 5G concepts to today’s 4G LTE 5.6 Potential throughput Calculated based on bandwidth, number of codewords and channel condition (only used for deactivation decision as CQI is only reported for activated SCells). 3CC DL Carrier Aggregation Extension L16B Enhancement The largest Transport Block Size (TBS) that can be carried with 100 PRBs and 256QAM is 97896 bits according to 3GPP TS 36.213. Since the Transmission Time Interval (TTI) is 1 msec this equates to a bit rate of 97.896 Mbps. Based on this TBS the peak throughput of one 20 MHz Cell Carrier (CC) is 97.896 X 2 = 195.792 Mbps assuming 2X2 MIMO and 97.896 X 4 = 391.584 Mbps, assuming 4X4 MIMO. In L16A the ‘3CC DL Carrier Aggregation Extension’ (FAJ 121 3084) optional feature supports a maximum of two layers per carrier and the ‘4×4 Quad Antenna Downlink Performance Package’ (FAJ 121 3076) does not support 4X4 MIMO with carrier aggregation and the ‘256 QAM Downlink’ (FAJ 121 4422) does not support more than two layers. - 154 - © Ericsson AB 2017 LZT1381950 R1A LTE Integrity KPIs Performance and related Parameters The peak downlink throughput using 3 CC carrier aggregation is 3 X 195.792 = 587.376 Mbps as illustrated in Figure 3-54. Largest TBS for 100 PRBs is 97896 bits 97896 bits in 1 msec = 97.896 Mbps 2X2 MIMO: 97.896 X 2 = 195.792 Mbps 4X4 MIMO: 97.896 X 4 = 391.584 Mbps L17A *UE limited to 10 layers L16B 20 MHz + 20 MHz + 256 QAM + 256 QAM + MIMO: = 195.792 = 391.584 MIMO: 256 20 MHz + QAM + = 195.792 20 MHz + 256 QAM + = 195.792 20 MHz + 256 QAM + = 195.792 20 MHz + 256 QAM + CAT 16 978.96 Mbps 20 MHz + 256 QAM + = 391.584 Figure 3-54: 4CC DL Carrier Aggregation Extension LZT1381950 R1A © Ericsson AB 2017 - 155 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop 5.7 Uplink Carrier Aggregation If we assume 6 out of the 100 uplink PRBs in a 20 MHz Cell carrier (CC) are allocated for PUCCH the largest Transport Block Size (TBS) that can be carried with the remaining 94 PRBs using 64 QAM is 68808 bits according to 3GPP TS 36.213. Since the Transmission Time Interval (TTI) is 1 msec this equates to a bit rate of 68.808 Mbps. Based on this TBS the peak uplink throughput of one 20 MHz Cell Carrier (CC) is 68.808 Mbps. 3GPP TS 36.213 V13.1.1 (2016-03) Table 7.1.7.2.1-1 Assuming 6 PRBs in 20 MHz Cell Carrier are reserved for PUCCH the largest TBS carried by remaining 94 PRBs = 68808 bits 68808 bits in 1 msec = 68.808 Mbps Carrier 1 Band A Carrier 2 Band B Carrier 1 Carrier 2 64 68.808 Mbps QAM Inter-Band 20 MHz + or or Intra-Band 137.616 Mbps Band C 20 MHz + Same Digital Unit Radio Node (DUS31/41) 64 68.808 Mbps QAM Baseband Radio Node (5216/5212) Figure 3-55: Uplink Carrier Aggregation The ‘Uplink Carrier Aggregation’ (FAJ 121 4425) L16B optional feature introduces support for uplink carrier aggregation of two inter-band noncontiguous or intra-band contiguous Cell Carrier combinations as specified in latest 3GPP 36.101. This means that the peak uplink throughput using carrier aggregation of two 20 MHz carriers assuming 64 QAM is used is 137.616 Mbps (68.808 X 2) as illustrated in Figure 3-55. Uplink Carrier Aggregation is not supported for intra-band non-contiguous carriers, inter-band 3 CC DL + 2 CC UL combinations that use two bands or FDD and TDD uplink combinations. UEs capable of 3 CC DL + 2 CC UL and 2 CC DL + 2 CC UL as well as UEs capable of just 2 CC DL + 2 CC UL are supported. The aggregated Cell Carriers must also belong to the same DUS31, DUS41, Baseband 5212 or Baseband 5216. 5.8 Uplink Carrier Aggregation Configuration Uplink Carrier Aggregation can only operate in conjunction with downlink carrier aggregation. The number of configured uplink aggregated carriers will always be equal to or less than the number of configured downlink aggregated carriers. - 156 - © Ericsson AB 2017 LZT1381950 R1A LTE Integrity KPIs Performance and related Parameters The UE will have one uplink Primary Cell (PCell) and at most one secondary cell (SCell) and will transmit the PUCCH in PCell only. SCells are first configured with an RRC Connection Reconfiguration message and then activated with a MAC Control Element. The total UE transmit power remains unchanged (23 dBm) independent of the number of UL component carriers. The UE receives UL grants per activated cell, and the HARQ feedback is transmitted in the same cell as where the grant was received. The UE controls on which component carrier the UL data is transmitted if it receives more than one grant in one TTI. eNodeB ManagedElement +-ENodeBFunction +-EUtranCellFDD/TDD networkSignallingValueCa = CA_NS_31 Enables additional spectrum emission requirements for CA according to 3GPP TS 36.101, table 6.2.4A-1. +-EUtranFreqRelation +-EUtranCellRelation sCellCandidate = ALLOWED The cell indicated by the value of parameter cellrelation can be used as an SCell CC for UEs that use this cell as their PCell CC for both UL and DL. +-CarrierAggregationFunction sCellActDeactUlDataThresh = 100 { 0..5000 } If the minimum time needed to transmit all bits in all priority queues in UL of a UE is higher than sCellActDeactUlDataThresh, activation of one or more secondary cells is considered. Unit: 0.1 Number of UL subframes sCellActDeactUlDataThreshHyst = 90 { 0..5000 } If minimum time needed to transmit all bits in all priority queues in UL of a UE is less than sCellActDeactUlDataThresh minus sCellActDeactUlDataThreshHyst, deactivation of one CC is considered. Unit: 0.1 Number of UL subframes Figure 3-56: Uplink Carrier Aggregation Parameters By setting the ‘sCellCandidate’ parameter illustrated in Figure 3-56 above to ‘ALLOWED’ the cell relation may be used as a SCell for both uplink and downlink Carrier Aggregation. As Multiple-Timing Alignment is not supported the distance between the antenna of the PCell and the antenna of the SCell must be less than 300m. In cases where the distance between the antenna of the PCell and the antenna of the SCell is greater than 300m the ‘sCellCandidate’ parameter should be set to ‘ONLY_ALLOWED_FOR_DL’ to allow downlink Carrier Aggregation only. The ‘NOT_ALLOWED’ and ‘AUTO’ settings for the ‘sCellCandidate’ parameter have the same effect of excluding the cell relation as a candidate for SCell for UEs using this cell as their primary component carrier (PCell). A UE is evaluated for Carrier Aggregation by checking its CA band and bandwidth combination support at initial attach, RRC connection reestablishment or incoming handover. LZT1381950 R1A © Ericsson AB 2017 - 157 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop An uplink CA Scell is selected from one of the downlink SCells so long as it is uplink Carrier Aggregation capable and its frequency is supported by the UE. If the minimum time needed to transmit all bits in all priority queues in UL of a UE is higher than the setting of the ‘sCellActDeactUlDataThresh’ parameter activation of a secondary cell is considered. The SCell is activated if either uplink or downlink activation criteria are met. The UE is first scheduled on the activated cell where it has the highest SINR. If it has enough power and buffer contents left after scheduling on the first cell it will be scheduled on the second cell as well. If minimum time needed to transmit all bits in all priority queues in UL of a UE is less than the value of the ‘sCellActDeactUlDataThresh’ parameter minus the ‘sCellActDeactUlDataThreshHyst’ parameter deactivation of the UL SCell is considered. The SCell is deactivated if both uplink and downlink deactivation criteria are met. The new ‘networkSignallingValueCa’ parameter illustrated in Figure 3-56 enables additional spectrum emission requirements for Carrier Aggregation according to 3GPP TS 36.101, table 6.2.4A-1 to be sent to the UE in the additionalSpectrumEmissionSCell-r10 Information Element. The default setting of ‘CA_NS_31’ according to 3GPP TS 36.101, table 6.2.4A-1 corresponds to no additional spectrum emission requirements. The ‘Carrier Aggregation’ (FAJ 121 3046/1) optional feature is a prerequisite for Uplink Carrier Aggregation and will enable a static 2 CC DL + 2 CC UL configuration. To support 3 CC DL CA + 2 CC UL CA the ‘3CC DL Carrier Aggregation Extension’ (FAJ 121 3084) and ‘Dynamic SCell Selection for Carrier Aggregation’ (FAJ 121 3063) optional features must be active. 5.9 Advanced Carrier Aggregation (FAJ 801 0568) The ‘Advanced Carrier Aggregation’ Value Package contains 3 functions. The ‘Inter-eNB Carrier Aggregation’ and ‘Carrier Aggregation-Aware Inter Frequency Load Balancing’ functions are enhanced in L17.Q1. Included Functions: › Inter-eNB Carrier Aggregation (Enhanced in L17.Q1) › Carrier Aggregation-Aware IFLB (Enhanced in L17.Q1) › Configurable SCell Priority Figure 3-57: Advanced Carrier Aggregation (FAJ 801 0568) - 158 - © Ericsson AB 2017 LZT1381950 R1A LTE Integrity KPIs Performance and related Parameters The ‘Inter-eNB Carrier Aggregation’ (FAJ 121 4469) optional feature makes it possible for Rel-10 Carrier Aggregation capable UEs to receive data from cells belonging to different eNodeBs. The eNodeB that has the serving cell (Master eNB) will forward user-data over the X2 interface to the external eNB that has the Scell (EScell). This feature makes it possible to use carrier aggregation in scenarios where the carriers are not deployed on the same eNodeB. The term ‘External Secondary Cell’ (ESCell) illustrated in Figure 3-58 below is used to denote a secondary cell that belongs to an external eNodeB. The greater flexibility in selection of secondary cells for carrier aggregation possible with the ‘Inter-eNB Carrier Aggregation’ optional feature allows the UE to use a nearly optimal set of cells for Carrier Aggregation thus improving end user experience in many scenarios. It also makes it easier to take advantage of under-utilized eNodeBs by making it easier for them to be used in carrier Aggregation scenarios. The Nodes involved in inter eNodeB Carrier Aggregation must be time and phase synchronized. This feature is only supported for FDD networks and only on the DUS31/41 hardware. The X2 one-way latency between master and external eNodeB must be lower than 9 ms, and for best performance 5 ms or less is preferred. This delay is constantly monitored by the ‘Inter-eNB Carrier Aggregation’ optional feature and when the maximum delay exceeds 9 ms, or the link quality is poor the inter eNodeB Carrier Aggregation activity is stopped between the affected eNodeB pair. The link will be periodically tested and Carrier Aggregation resumed between the pair when appropriate. Carrier Aggregation possible in scenarios where the carriers are not deployed on the same eNodeB External eNodeB Support in 52 series hardware added in L17.Q1 FDD PCell + TDD SCell supported in L17.Q1 52xx 52xx Master Master eNodeB eNB X2 S 1 Small eNodeB 52xx 52xx Macro eNodeB X2 UEs in coverage area of Small cell can now perform Carrier Aggregation with Macro cell Figure 3-58: Inter-eNB Carrier Aggregation Another typical use-case illustrated in Figure 3-58 above is where a small cell is deployed in the coverage area of a macro but supporting fewer frequencies than the macro. Without this feature UEs in the coverage area of the small cell could not perform carrier aggregation with the frequencies of the macro cell however with this feature it is now possible. LZT1381950 R1A © Ericsson AB 2017 - 159 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop 5.9.1 Inter-eNB Carrier Aggregation Performance The inter-eNodeB Carrier Aggregation performance is dependent on the X2 and processing delays. In the illustration here we can see how these delays affect the standard 8 TTI HARQ loop meaning that the UE cannot be scheduled every TTI. • To prevent control channel scheduling conflict with local UEs the Modulation Coding Scheme is limited to 27 for inter eNodeB traffic. This means that inter-eNodeB Carrier Aggregation cannot achieve the same peak bit rates as intra-EnodeB or Elastic RAN Carrier Aggregation. • The graph shown here illustrates the peak throughput Efficiency Fraction for an External SCell with 20 MHz Bandwidth for X2 latencies between 1 and 9 msec. Scheduling at eNodeB Retransmission in response to 1N (NACK) Non-ESCell-UE HARQ response timing loop UE ACK (A) or NACK (N) UE cannot be scheduled every TTI in ESCell Scheduling at eNodeB ESCell-UE HARQ response timing loop 3GPP TS 36.213 V13.1.1 Table 7.1.7.2.1-1 To prevent control channel scheduling conflict with local UEs the Modulation Coding Scheme is limited to 27 for inter eNodeB traffic. 20 MHz Bandwidth ESCell Efficiency Fraction UE ACK (A) or NACK (N) X2 Latency (msec) Figure 3-59: Inter-eNB Carrier Aggregation Performance - 160 - © Ericsson AB 2017 LZT1381950 R1A LTE Integrity KPIs Performance and related Parameters 5.9.2 Inter-eNB Carrier Aggregation Parameters The configuration process for inter eNodeB Carrier Aggregation is very similar to legacy Carrier Aggregation and makes it possible to set the ‘sCellCandidate’ parameter to ‘ALLOWED’ or ‘ONLY_ALLOWED_FOR_DL’ for up to 12 ‘ExternalEUtranCellFDD’ MOs as illustrated in Figure 3-60 below. External cells will automatically be excluded by the ‘Uplink Carrier Aggregation’ optional feature so setting the ‘sCellCandidate’ parameter to ‘ALLOWED’ or ‘ONLY_ALLOWED_FOR_DL’ is effectively the same. eNodeB ManagedElement +-ENodeBFunction +-EUtranCellFDD +-EUtranFreqRelation +-EUtranCellRelation EUtranCellRelationId = ExternalEUtranCellFDDId sCellCandidate = ALLOWED or ONLY_ALLOWED_FOR_DL esCellCaConfigurationAvail = ENABLED If indicated cell is currently able to serve as ESCell. New read-only parameter +-EUtraNetwork +-ExternalENodeBFunction +-TermPointToENB New read-only parameter interEnbSwCompatibilityState = COMPATIBLE Reflects whether the eNodeBs designated in the endpoints between eNodeBs have inter-eNodeB-compatible capabilities. +-TermPointToLbm New MO operationalState TermPointToLbmId +-ExternalEUtranCellFDD New read-only parameter remoteCellState = ENABLED Reflects whether the indicated cell is currently able to serve as an SCell. Figure 3-60: Inter-eNB Carrier Aggregation Parameters The new 'termPointToLbm' MO class illustrated in Figure 3-60 above, is created by the system when an eNB establishes communication between its own and a partner's LBM (LTE baseband modules) for the purposes of Inter-eNB Carrier Aggregation. The need for Inter-eNB Carrier Aggregation configuration is triggered when the 'sCellCandidate' parameter is set to 'ALLOWED' or 'ONLY_ALLOWED_FOR_DL' and the referenced SCell is external. The ‘esCellCaConfigurationAvail’, ‘interEnbSwCompatibilityState’ and ‘remoteCellState’ read-only parameters illustrated in Figure 3-60 above are also introduced with the ‘Inter eNodeB Carrier Aggregation’ optional feature. LZT1381950 R1A © Ericsson AB 2017 - 161 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop 5.10 Multi-carrier Load Management (FAJ 801 0424) The 'Multi-carrier Load Management' Value Package contains 6 functions. The 'Inter-Frequency Load Balancing' and 'UE Throughput-Aware IFLB' functions are enhanced in L17.Q1. Included Functions: › Best Neighbor Relations for Intra-LTE Load Management › Coverage-Adapted Load Management › IFLB Activation Threshold › Inter-Frequency Load Balancing (Enhanced in L17.Q1) › Limited Uplink-Aware IFLB › UE Throughput-Aware IFLB (Enhanced in L17.Q1) Figure 3-61: Multi-carrier Load Management (FAJ 801 0424) 5.11 Prescheduling Prescheduling is a feature that minimizes the uplink round trip time by blindly giving PUSCH grants to UE in advance. This is achieved by sending UL scheduling grants without explicit scheduling request from the UE side. Prescheduling grants are only sent in low load. › Prescheduling is a method to minimize UL delay (and hence round trip time) by means of blindly giving PUSCH grants to a UE in advance, without receiving buffer status information (SR, BSR). Scheduling request is no longer mandatory to get UL grant SR Grant Figure 3-62: Prescheduling Prescheduling is used to improve user plane and initial control plane latency which leads to the following enhanced areas: • Reduced download time, especially for complex web pages where the improvement is significant. • Clear improvement of the so-called speed-tests (pingtests). Different versions of speed-test are commonly used by smart phones users to measure bandwidth quality. - 162 - © Ericsson AB 2017 LZT1381950 R1A LTE Integrity KPIs Performance and related Parameters • Faster connection set-up times. The Operator benefits are: • Reduced End-user latency • Low latency is important for small data transmission and is typically measured by broadband network test applications Pre-scheduling is persistently allocating UL transmission resources to the device, therefore User plane latency is reduced due to less signaling. Pre-scheduling allocates transmission opportunities in the uplink even if no scheduling request has been received from the device. The device has to respond even if there is no data to send. An example of a Round Trip Delay measurement with prescheduling OFF is illustrated in Figure 3-63 below. Delay Breakdown PING 0 SR encod. and alignment 5 5 SR 1 PING 15 20 PING eNB processing 3 Grant 1 10 3GPP specified delay 4 1 PING PING 1 3 PING Core network delay 2 1 Time [ms] 1 23 ms Total round trip delay * *Note: Numbers are example only Figure 3-63: Prescheduling OFF LZT1381950 R1A © Ericsson AB 2017 - 163 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop An example of a Round Trip Delay measurement with prescheduling ON is illustrated in Figure 3-64 below. Delay Breakdown PING 0 Grant PING 5 10 PING 3GPP specified delay 4 1 PING PING 1 3 PING Core network delay 2 1 1 15 20 Time [ms] 13 ms Total round trip delay * *Note: Numbers are example only Figure 3-64: Prescheduling ON 5.11.1 Prescheduling Parameters The parameters that control the Prescheduling feature are illustrated in Figure 365 below. ManagedElement +-ENodeBFunction initPreschedulingEnable Indicates that prescheduling is enabled during connection setup phase +-PredchedulingProfile preschedulingDataSize Granted data size during prescheduling period. preschedulingDuration Prescheduling is stopped if there are no UL transmissions during this time. preschedulingPeriod Period in ms lapsed between granted prescheduling instances Figure 3-65: Prescheduling Parameters - 164 - © Ericsson AB 2017 LZT1381950 R1A LTE Integrity KPIs Performance and related Parameters 5.12 LTE transmission modes Downlink transmission modes The LTE transmission modes for downlink is shown in Figure 3-66. The 4x4 Quad Antenna DL Performance Package provides support for up to rank 4 DL transmission for UEs that support it (UE Category 5). This enables 4-way to transmit diversity and 4-way spatial multiplexing. It also results in improvements to downlink coverage and capacity, especially in high-quality channel conditions. This gives twice the peak rate compared to rank 2. TM3 (open loop spatial multiplexing) as well as TM4 (closed loop spatial multiplexing) is supported for both FDD and TDD. With this feature operator will be able to configure the 4TX cells with either TM3 or TM4 for both TDD and FDD. Dynamic switching between the modes may be implemented later. For TDD, it will support TDD ul-dl subframe configuration 2, and special subframe configurations 6 and 7. Supported 4-layer capable UE category is 5. This feature uses existing Key Performance Indicators (KPIs) for system performance evaluation and monitoring. Significant DL throughput gain is expected with this feature compared to 4x2 MIMO. It is preferable to measure the PM counters in the network having high traffic activity, and compare calculated DL cell throughput to the performance achieved by 4X2 MIMO under the same conditions. Other DL throughput related KPIs such as DL PDCP UE throughput can also be used to measure DL throughput gain. There should be significant improvement on other KPIs as well compared to 4x2 MIMO. The feature is license controlled. The licensing MO instance name is ‘QuadAntDlPerfPkg4x4’. As a prerequisite, the following feature shall be enabled: • 4x2 Quad Antenna Downlink Performance Package, FAJ 221 3041. For TDD, it will support TDD ul-dl subframe configuration 2, and special subframe configurations 6 and 7. In both FDD and TDD, this feature requires the following RAN features to be active: • Dual-Antenna Downlink Performance Package • 4x2 Quad Antenna Downlink Performance Package This feature cannot be enabled if the following feature is active: • TM9 8x2 Octal Antenna Downlink Performance Package • Combined Cell • Octal Antenna Uplink Performance Package LZT1381950 R1A © Ericsson AB 2017 - 165 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop The recommended antenna configuration depends on the extent that 4x4 MIMO capable UEs are deployed in the network compared to 4x2 MIMO only capable UEs. • When there are relatively few 4x4 MIMO capable UE’s in the cell, two columns of closely spaced cross-polarized antenna elements are recommended in order to maximize beam forming gain. • When there are relatively many 4x4 MIMO capable UE’s in the cell, two spatially separated columns of cross-polarized antenna elements are recommended in order to maximize the rank of the RF channel. There is a fundamental trade-off between maximizing beamforming gain versus maximizing the rank of the RF channel. The former improves cell edge performance for 4x2 MIMO capable UEs. The latter improves peak throughput for 4x4 MIMO capable UEs only. When switching from 2TX to 4TX deployment: • It is necessary to re-tune the ‘cfiMode’, CRS scaling, HO thresholds (‘qRxLevMin’, A2/A5/B2 thresholds) settings that may have been optimized for 2TX but are not necessarily optimum for 4TX. • For the same cell power and CRS scaling, the measured RSRP at a given location in the cell configured will be lower by 3dB going from 2TX to 4TX. This is due to the way RSRP is defined by 3GPP standard. The change in RSRP levels is what necessitates re-tuning of HO thresholds. • Consider setting ‘cfiMode’ to 4 or 5. This helps mitigate the fact that there are fewer CCEs for PDCCH with 4TX compared to 2TX since the introduction of 2 additional CRS comes at the expense of REs available for PDCCH and PDSCH. The Adjustable CRS Power feature sets default values for two parameters, ‘crsGaincrsGain’ and ‘pdschTypeBGain’. However, these default values are not always suitable for 4TX MIMO cells of large size. In large cells, these default values must be tuned to maximize the performance for feature ‘4×4 Quad Antenna Downlink Performance Package’. If the cell is coverage-limited, the parameters are tuned based on network deployment density. If the cell is interference-limited, the parameters are tuned based on network load. In release 11, enhanced support for ‘CoMP’ (Co-ordinated Multipoint Transmission/reception) with transmission mode 10 is introduced. - 166 - © Ericsson AB 2017 LZT1381950 R1A LTE Integrity KPIs Performance and related Parameters Different MIMO related features are discussed in ‘LTE L17 Advanced Radio Network Features’ course. Rel-8 Rel-9 › TM1 “Single-antenna port” › TM2 “Transmit diversity” › TM3 “Open-loop spatial multiplexing” › TM4 “Closed-loop spatial multiplexing” › TM5 “Codebook based MU-MIMO” › TM6 “Rank-1 closed-loop spatial multiplexing” › TM7 “Single layer transmission” › TM8 “Dual-layer transmission” 1, 2 or 4 antenna ports with cell-specific reference signals (CRS) UE-specific reference signals for demodulation (DM-RS) Rel-10 › TM9 “Up-to-8 layer transmission” + 1 CSI-RS resource for enhanced feedback Rel-11 › TM10 “TM9 + CoMP support” + multiple CSI-RS resources, CSI-IM – new interference measurements, antenna port co-location, PDSCH mapping flexible configurability of RS sequences Figure 3-66: LTE transmission modes Downlink transmission modes 5.12.1 Which MIMO Mode is Best? The best transmission mode to use will depend on the antenna configuration as summarized in Figure 3-67. › 2TX (Crosspol): – Best mode is TM3 (OLSM) – There is little or no benefit in running TM4 (CLSM) for 2TX. › 4TX (Dual crosspol): – Best mode for 4TX radio is TM4 (CLSM) (4x4,4x2). › 8TX – TM8 is today most efficient. However TM8 was introduced in Rel-9, thus not all terminals supports TM8. Then TM7 or TM3 will be used – TM8 is as TM7 only supported in TDD – TM9 offers higher capacity and cell edge bit rates – TM9 will be applicable for both FDD and TDD – In order to run 8 layers spatial multiplexing TM9 is needed Figure 3-67: Which MIMO Mode is Best? LZT1381950 R1A © Ericsson AB 2017 - 167 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop VOLTE Impact on BB Throughput & DBS 6 The VoIP co-exist together with Packet Switched (PS) data in a network. PDCCH can be the limiting resource when VoIP traffic is present in the system. This is not normally the case for other PS services. If there are large numbers of VoIP users, other PS services are blocked and capacity is not fully used. PS throughput will be zero when the system reaches the PDCCH limitation. VoIP is a GBR service, which means that it has priority in the scheduler before other PS services like mobile broadband. VoIP is given QCI = 1 marking it as GBR. The introduction of DBS and SABE is expected to increase the capacity in mixed scenarios: • Increased throughput for PS users for the same amount of VoIP users in a cell • Increased number of VoIP users per cell for the same PS throughput The Delay-Based Scheduling and Grant Estimation feature provides improved average Uplink (UL) and Downlink (DL) throughput for Mobile Broadband (MBB) services in a mixed scenario with both Voice over IP (VoIP) and MBB services while not reducing the VoIP capacity, specifically when a large number of VoIP users are active in the cell. It is also possible to increase the VoIP capacity instead of improving the average UL and DL throughput for MBB services. › RR: Round Robin Scheduler with QoS aware scheduling, VoIP users will always have higher priority than Data Users. › DBS: Delay based Scheduling and Grant Estimation used. Delay based scheduling gains seen in at high VoIP load, where the scheduling capacity is limited Figure 3-68: VOLTE Impact on BB Throughput & DBS By applying an age-dependent weight to VoIP packets, scheduling resources are only used for VoIP services when this is required to avoid exceeding the ‘Packet Delay Budget (PDB)’. This results in the scheduling of several VoIP packets during a scheduling instance and efficient use of the RBS resources. Figure 3-68 and Figure 3-69 present the simulation results with the feature Delay Based Scheduling on and without the feature (Round Robin used instead). - 168 - © Ericsson AB 2017 LZT1381950 R1A LTE Integrity KPIs Performance and related Parameters Figure 3-69: VOLTE Impact on BB Throughput & DBS When the ‘schedulingAlgorithm’ parameter for a particular QCI is set to ‘DELAY_BASED’ as illustrated in the Figure 3-70 for QCI 1, the scheduler will give it a small scheduling weight before what is known as the ‘bundling time’. This weight will slowly increase as the age of the oldest packet in the buffer increases. Once the age of the oldest packet reaches the ‘bundling time’ the scheduling weight will be sharply increased. Although this will not guarantee that the data is scheduled it will greatly improve its chances. At low load the data could be scheduled before bundling time but at high load it may be scheduled after the ‘bundling’ time. The ‘bundling time’ is hardcoded to 40 msec. Scheduling weight Larger Weight after bundling time, but not guaranteed scheduling eNodeB ManagedElement +-ENodeBFunction +-QciTable +-QciProfilePredefined=qci1 schedulingAlgorithm = DELAY_BASED Small weight increase before reaching the bundling time Bundling Time (Hard coded to 40 msec) Age of oldest packet Figure 3-70: VOLTE Impact on BB Throughput & DBS LZT1381950 R1A © Ericsson AB 2017 - 169 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop Summary 7 The participants should now be able to: 3 Explain LTE Integrity KPIs performance and related parameters 3.1 Describe different LTE integrity KPIs & counters related to Throughput, Latency 3.2 Explain associated parameters related to different Integrity KPIs 3.3 Relate QOS and Scheduling profile for integrity optimization 3.4 Analyze steps to effectible troubleshoot throughput issues 3.5 Explore related features and associated parameters for throughput enhancement 3.6 Measure VOLTE effect on throughput and data services KPIs Figure 3-71: Summary of Chapter 3 - 170 - © Ericsson AB 2017 LZT1381950 R1A Issue Analysis, Improvements and Case-Studies for Integrity KPIs 4 Issue Analysis, Improvements and CaseStudies for Integrity KPIs Objectives After completion of this chapter the participants will be able to: 4 4.1 4.2 4.3 4.4 Classify the issues, analyze them for improvements and discuss few case studies for Integrity KPIs. Identify steps for UL/DL Throughput investigation & essential check Describe DL & UL throughput optimization strategy Discuss different cases of throughput & Latency degradation Analyze cases based on low throughput high latency exercises and provable solutions Figure 4-1: Objective of Chapter 4 LZT1381950 R1A © Ericsson AB 2017 - 171 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop Case-1: DL Throughput improvement 1 A Network has been observed to have low downlink average throughput, but while changing the scheduling algorithm from proportional fair high to proportional fair low, the average cell throughput has improved by around 30%. Cell capacity also shows significant improvement, but the algorithm has reduced cell edge user throughput. Hence, for throughput KPI optimization the correct scheduling algorithm does play an important role. In many scenarios, cell capacity can be increased by using the low fairness versions of the proportional fair scheduling algorithms. These algorithms increase the share of the resources given to users with good channel conditions. This leads to an overall increase in capacity. The minimum rate feature helps in giving a consistent cell edge throughput with the minimum rate Quality of Service characteristic. The basic concept is that flows experiencing a bit rate lower than their configured minimum bit rate will be prioritized before flows that have a bit rate above their configured minimum rate. As mentioned above, Proportional Fair scheduling feature takes into account both scheduled data rate and radio channel quality. The tradeoff between user fairness and the system performance can be tuned by Channel Quality Fraction (CQF). The CQF controls how big portion of channel quality should contribute to a user's priority. OBJECTIVE: To check Impact on Dl Avg Sector Throughput by changing Scheduling Algorithm. Proportional Fair(High) to Proportional Fair Low. 1.Proportional Fair High(2): Pros and Cons? Pros- Each user Treat Equally. Cons- User which is in Good RF Condition not get RB as per Desired.Finally Overall Average Sector/UE Throughput Dip . 2. Proportional Fair Low(0): Pros and Cons? Pros- Avg sector/UE Throughput improved (30-40)% Theoretically while at the same time cell edge user get at least Minimum bit rate. Capacity improved. Cons- Less fair towards Cell edge user. Figure 4-2: CASE 1: DL Throughput - 172 - © Ericsson AB 2017 LZT1381950 R1A Issue Analysis, Improvements and Case-Studies for Integrity KPIs 2 Case-2 UL-Throughput improvement An Operator has observed a low UL throughput; further observations also indicates towards a poor throughput in few cells in some areas of the network. The throughput plot shows the low throughput cells and areas, as shown in Figure 4-3. After checking the QOS parameters and some other investigations the optimization team has suggested the following parameter changes. ‘pZeroNominalPucch’: From -117 to -103 ‘pZeroNominalPusch’: From -106 to -96. P0_PUSCH is the target PSDrx, for each resource block, set according to the parameter ‘pZeroNominalPusch’. It is common for all UEs in the cell ‘P0_PUCCH’ is the target PSDrx, corresponding to ‘P0_PUCCH’ from PUSCH. It is set according to the parameter ‘pZeroNominalPucch’ and signaled separately on BCCH System Information Blocks (SIBs) Increasing the values of the ‘pZeroNominalPusch’ and ‘pZeroNominalPucch’ parameters result in higher signal power, but will also result in increased intercell interference, which may affect cell edge UE. Decreasing the values of the parameters will have the opposite effect. With parameter change ‘pZeroNominalPucch’ from -117 to -103 and ‘pZeroNominalPusch’ from -106 to -96 gave good improvement for UL throughput, the UL-throughput plot is shown in Figure 4-3 below. An Operator complain about low UL throughput, stats shows the low throughput in some cells. Changed the parameter and get good improvement in UL throughput. pZeroNominalPucch: From -117 to -103 pZeroNominalPusch: From -106 to -96. What may be negative effect of this? Post Pre Figure 4-3: CASE 2: UL Throughput Improvement LZT1381950 R1A © Ericsson AB 2017 - 173 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop With fractional path-loss power control, an operator configurable cell specific pathloss compensation factor (α) is introduced. Alpha < 1.0, allows users to be received with variable PSD depending on their path loss i.e. the user with small path loss will be received with high PSD. This parameter is broadcast in the cell in SIB2. The broadcast alpha is used for Semi Persistent Scheduling and dynamic scheduling, while ‘alpha’ = 1 is used for RA response. With fractional path-loss power control, it is easier to set power control parameters that will enable UL peak rate in the cell without sacrificing cell edge performance too much. Besides, trade-off will be possible between average and cell edge throughput. The benefit of the features can be observed via the uplink user throughput distribution taken over a sufficiently long period of time. With fractional pathloss power control, the trade-off between peak, average and cell-edge throughput levels should be achieved with a combined setting of the parameters, ‘pZeroNominalPusch’ and ‘alpha’. By varying the value of alpha for a given ‘pZeroNominalPusch’ (P0) value. Uplink received power and user throughput is reduced when lowering alpha, given the interference level is constant. Thus, a lower ‘alpha’ should be accompanied by a higher P0 to maintain throughput at the cell edge Case-3: LTE Latency Issue 3 For a Network the overall latency degraded. Overall downlink latency and latency per QCI in Figure 4-4 below shows, the latency of QCI1 degraded most. The overall latency degraded and the latency of QCI1 degraded most Figure 4-4: Case 3: LTE Latency issue - 174 - © Ericsson AB 2017 LZT1381950 R1A Issue Analysis, Improvements and Case-Studies for Integrity KPIs 3.1 Case-3: Investigation analysis Parameter audit and the investigation of recent network changes, followed by the configuration logs were checked for the said duration, while during the degradation. Configuration Log investigation shows the major reason is that the feature ‘SS DRX’ was activated during the time, Figure 4-5. › The degradation occurred around 17th Feb to 23rd Feb › Configuration Log investigation shows the major reason is feature SS DRX Before After 2 questions were submitted: 1. Why did SS DRX impact the Latency ? 2. Why was QCI1 impacted most ? Figure 4-5: Case 3: Investigation Analysis ‘Service-specific DRX’ feature allows the operator to configure, default connected-mode DRX parameters and service-specific connected mode DRX parameters. Service specific connected mode DRX parameters refer to DRX settings that change depending on the services that are established. This is based on the QCI values of the bearers. The connected mode DRX parameters can be adjusted based on the services that are currently in use by the UE. Service Specific DRX Introduction SS DRX Overview: Service-specific DRX allows the operator to configure the following: • Default connected-mode DRX parameters • Service-specific connected mode DRX parameters SS DRX Benefit: The connected mode DRX parameters can be adjusted based on the services that are currently in use by the UE. • This feature allows a mobile broadband connected mode DRX configuration to be used by default, but to switch to a VoLTE-specific DRX value when a VoLTE call is established. • Significant Improvement to Battery Saving during a Voice Call. Figure 4-6: Case 3: Investigation Analysis LZT1381950 R1A © Ericsson AB 2017 - 175 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop Service-specific connected mode DRX parameters can be considered to be redefined. Instead of having one setting for all the services, a set of DRX settings can be configured by operator, providing the “best fit” DRX settings for certain services. DRX configuration is chosen based on current UE’s primary service configured by the operator. Each QCI associates one DRX profile, based on its services characteristics. The same DRX profiles can be set to multiple QCIs. Each QCI has ‘drxPriority’, a higher value indicates a higher priority for the associated DRX profile if multiple QoS are serving a UE. Among QCIs, ‘drxPriority’ should be set with a unique value unless the same DRX Profile is associated for those QCIs. If two QCIs are configured with the same ‘drxPriority’ values but different ‘drxProfiles’, eNB could randomly pick one of the profiles for DRX settings. DRX configuration for ANR CGI measurement takes a higher priority than QCI based DRX configurations 3.2 Case-3 Investigation for Service Specific DRX The service specific DRX allowed the oprator to set short DRX cycle and long DRX cycle. A packet can arrive at the eNB when the UE is in SLEEP mode This means that the packet will be stored in the eNB buffer until the UE is in WAKE mode and this causes Extra Latency. DRX is a balance between latency and battery saving. Figure 4-7 below illustrate the function of SS DRX. Why SS DRX impact the Latency › A packet can arrive at the eNB when the UE is in SLEEP mode › This means that the packet will be stored in the eNB buffer until the UE is in WAKE mode › This causes Extra Latency › DRX is a balance between latency and battery saving Extra latency time eNB receives packet for UE, but stores it in the buffer until UE had finished DRX cycle SLEEP longDRX-Cycle CycleStartO ffse UE monitors the PDCCH One successful decoding of PDCCH frame Short DRX is only for one cycle before Starts the inactivity time going to long DRX shortDrxCycleTimer =1 over the air Ue Mode DurationTimer = 10ms WAKE Packet transmitted Packet held in buffer eNB Buffer drx-inactivity timer shortDRX- longDRX-Cycle UE monitors the PDCCH UE monitors the PDCCH Cycle UE monitors the PDCCH Figure 4-7: Case 3: Investigation Analysis 3.3 Case-3 DRX Profile Parameter Optimization The UE specific DRX and the ‘ServiceSpecificDrx’ are optional features, with both the feature licenses, the operator is able to configure an extended DRX profiles table and can map a DRX profile to QCI and set DRX priority. - 176 - © Ericsson AB 2017 LZT1381950 R1A Issue Analysis, Improvements and Case-Studies for Integrity KPIs Up to 19 DRX profiles can be configured by operator, each profile contains two DRX configurations for UEs supporting or not supporting short DRX respectively. Under the same DRX profile, ‘longDrxCycleOnly’ does not have to be the same as ‘longDrxCycle’, so advantage of UE’s capability on supporting shortDrxCycle can be well taken. ‘QciProfileOperatorDefined’ (or Predefined) ‘drxPriority’, the relative priority among the DRX profiles, i.e. if the bearer that is setup with this QCI has a higher DRX priority than any of the existing bearers, the DRX configuration will be set to those selected by the ‘drxProfileId’ for this QCI. The ‘drxPriority’ has to be unique across all the configured ‘QciProfilePredefined’ and ‘QciProfileOperatorDefined’ MOC instances except for instances where the drxProfileId is the same. That is, instances that share the same ‘drxProfileId’ may have the same ‘drxPriority’ value. Also note that larger drxPriority values indicate higher relative priority. Change takes effect: Object unlocked. Range: 0 to 100, Default=0 In this case QCI 1 has highest ‘drxPriority’= ‘10’ and the other parameters set for QCI 1 is not correctly optimized. Let’s analyze the set parameters for QCI1 shown in Figure 4-8. The value of ‘longDrxCycle’ is in number of sub-frames. If ‘shortDrxCycle’ is configured, the value of ‘longDrxCycle’ shall be a multiple of the ‘shortDrxCycle’ value. The ‘longDrxCycleOnly’ value is used as ‘longDrxCycle’ when the UE does not support ‘shortDrxCycle’. The value of ‘longDrxCycleOnly’ is in number of sub-frames. Why QCI1 was impacted most Current setting › The QCI 1 DRX profile need to be optimized. Figure 4-8: Case 3: Investigation Analysis The parameters can be understood below as: • ‘DrxProfile’-‘drxRetransmissionTimer’: Specifies the maximum number of consecutive PDCCH-subframe(s) as soon as a DL LZT1381950 R1A © Ericsson AB 2017 - 177 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop retransmission is expected by the UE. TS36.331 Ch. 6.3.2, RRC parameter ‘drx-RetransmissionTimer’. Change takes effect: Object unlocked, Default=PSF2, Range: 0,1,2,3,4,5,6,7 • ‘DrxProfile’-‘shortDrxCycle’: Specifies the consecutive subframes between on-duration phases, i.e. the period of re-occuring onduration phases. The parameter is applied when the UE follows the short DRX cycle. The parameter is ignored if ‘shortDrxCycleTimer’ = 0. In that case ‘longDrxCycle’ is applied and the UE follows the long DRX cycle only. Corresponds to 3GPP information element ‘shortDRX-Cycle’. Dependencies: The use of ‘shortDrxCycles’ is disabled if the attribute shortDrxCycleTimer = 0. Change takes effect: Object unlocked, Default=SF40, Range: 0,1,2,3,4,5, 6…,15 • DrxProfile-‘shortDrxCycleTimer-long’: Specifies the number of consecutive subframes the UE must follow the short DRX cycle after the DRX Inactivity Timer has expired. Value in multiples of shortDRX-Cycles. A value of 1 corresponds to 1 * ‘shortDRXCycle’, a value of 2 corresponds to 2 * ‘shortDRX-Cycle’ and so on. Corresponds to 3GPP timer drxShortCycleTimer. The value 0 means that short DRX cycles are not used. Change takes effect: Object unlocked, Range: 0 to 16, Default=4 Case-Study 4: Low throughput during High Load 4 The customer has reported that on 5 Mhz sites which cover special events, the users have no significant throughput when there are high number of RRC connected users in the cell. - 178 - © Ericsson AB 2017 LZT1381950 R1A Issue Analysis, Improvements and Case-Studies for Integrity KPIs The Scheduling algorithm is set to Proportional Fair Low. This feature makes it possible to benefit from the fact that different users experience different radio conditions at a certain time. By prioritizing users experiencing good channel quality, a higher throughput can be achieved, compared to the Resource Fair algorithm. However, to avoid some users being allocated too few or no resources due to poor channel quality, fairness is provided by also considering the average rate in scheduler prioritization. › Case Details: Special event sites with 5 MHz Scheduling algorithm is set to Proportional Fair Low › Proportional Fair Low: – Benefit from the fact that different users experience different radio conditions at a certain time. – A higher throughput can be achieved when compared to Resource Fair algorithm – It prioritizes users experiencing good channel quality – To avoid some users being allocated too few or no resources due to poor channel quality, fairness is provided by taking into account the average rate in scheduler prioritization Figure 4-9: Case-Study 4: Reported issue 4.1 Exercise 1 – Analyze ROP files Collected counters (ROP file excel format) are provided by instructor and kept at \Exercises\ROP_files\ Exercise 1: Analyze the ROP files in order to determine why low throughput is experienced when a high number of users are connected to the cell. Analyze the ROP files in excel format given by instructor, in order to find the cell and time period where low downlink throughput is observed. Hints: Use the KPI files (given excel files) to get an overview of the downlink throughput. LZT1381950 R1A © Ericsson AB 2017 - 179 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop Figure 4-10 is a brief introduction for exercise 1. › Collected data: ROP files: \Exercises\ILL00081_ropfiles\ILL00081_ropfiles › Analyze the ROP files in order to determine why low throughput is experienced when a high number of users are connected to the cell. › Use the ROP files (excel) in order to find the cell and time period where low downlink throughput is observed. › Hints: – Use the chart option in the provided KPI excel to get an overview of the downlink throughput. Figure 4-10: Exercise 1 – Analyze ROP files 4.1.1 Exercise 1 – Answer A Cell ILL00081_7A_1 starts to experience roughly no throughput at 10-28 16:30 PM and lasts tills 10-28 20:15 PM which is roughly 4 hours’ duration. › Cell ILL00081_7A_1 starts to experience roughly no throughput at 28 Oct 16:30 and lasts tills 28 Oct 20:15 which is roughly 4 hours. Figure 4-11: Exercise 1 - Answer 4.2 Exercise 2: Possible cause for throughput degradation Once you have identified the cell and time-period when the problem starts to occur then we can move onto determining the possible causes for throughput degradation. - 180 - © Ericsson AB 2017 LZT1381950 R1A Issue Analysis, Improvements and Case-Studies for Integrity KPIs The next step is to list counters which can indicate possible bottlenecks to validate. Hints: In order to have good throughput; the RF environment needs to be good as well. List counters which can be used to validate the RF environment › Once you have identified the cell and time period when the problem starts to occur then we can move onto determining the possible causes for throughput degradation. › The next step is to List counters which can indicate possible bottlenecks to validate. › Hints: In order to have good throughput then the RF environment needs to be good as well. List counters which can be used to validate the RF environment? Figure 4-12: Exercise 2: 4.2.1 Exercise 2 – Answer UL/DL poor RF counters Name some counters related to UL and DL interference in Figure 4-13 › UL Interference: pmRadioRecInterferencePwr : The measured Noise and Interference Power on PUSCH, › DL Interference: pmRadioUeRepCQIDistr: Rank 1 (TX Diversity) pmRadioUeRepCQIDistr2: Rank 2 (MIMO) Figure 4-13: Exercise 2 - Answer LZT1381950 R1A © Ericsson AB 2017 - 181 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop 4.2.2 UL Noise and Interference The Power on PUSCH can be observed with pmRadioRecInterferencePwr’ which gives the measured noise and interference power on PUSCH, according to 36.214. › UL Interference: pmRadioRecInterferencePwr : The measured Noise and Interference Power on PUSCH, according to 36.214. PDF ranges: [0]: N+I <= -121 [1]: -121 < N+I <= -120 [2]: -120 < N+I <= -119 [3]: -119 < N+I <= -118 [4]: -118 < N+I <= -117 [5]: -117 < N+I <= -116 [6]: -116 < N+I <= -115 [7]: -115 < N+I <= -114 [8]: -114< N+I <= -113 [9]: -113 < N+I <= -112 [10]: -112 < N+I <= -108 [11]: -108 < N+I <= -104 [12]: -104 < N+I <= -100 [13]: -100 < N+I <= -96 [14]: -96 < N+I <= -92 [15]: -92 < N+I Figure 4-14: Exercise 2 – Answer DL Interference can be measured with ‘pmRadioUeRepCQIDistr’: Rank 1 (TX Diversity) ‘pmRadioUeRepCQIDistr2’: Rank 2 (MIMO), PDF counters, Figure 4-15. › DL Interference: pmRadioUeRepCQIDistr: Rank 1 (TX Diversity) pmRadioUeRepCQIDistr2: Rank 2 (MIMO) PDF ranges: [0]: CQI = 0 [1]: CQI = 1 [2]: CQI = 2 [3]: CQI = 3 [4]: CQI = 4 [5]: CQI = 5 [6]: CQI = 6 [7]: CQI = 7 [8]: CQI = 8 [9]: CQI = 9 [10]: CQI = 10 [11]: CQI = 11 [12]: CQI = 12 [13]: CQI = 13 [14]: CQI = 14 [15]: CQI = 15 Figure 4-15: Exercise 2 – Answer cont’d - 182 - © Ericsson AB 2017 LZT1381950 R1A Issue Analysis, Improvements and Case-Studies for Integrity KPIs 4.2.3 PDCP Layer counters Name some counters, provide the number of packets received, transmitted, and discarded at the PDCP layer. Write down the answers in Figure 4-16 below. ➢PDCP Layer counters: The following counters will provide the number of packets received, transmitted, and discarded at the PDCP layer pmPdcpPktTransDl pmPdcpPktReceivedDl Figure 4-16: Exercise 2 – Answer cont’d pmPdcpPktDiscDlPelr pmPdcpPktDiscDlPelrUu 4.2.4 RLC Layer counters The counters below provide an indication for the total number of successful and unsuccessful transmissions at the RLC layer. ‘pmRlcArqDlAck’, ‘pmRlcArqDlNack’, ‘pmRlcArqUlAck’, ‘ pmRlcArqUlNack’, Figure 4-17. Following formulas can be used to determine the success rate at RLC.RLC UL success rate % = pmRlcArqUlAck/( pmRlcArqUlAck + pmRlcArqUlNack) * 100% RLC DL success rate % = pmRlcArqDlAck/(pmRlcArqDlAck +pmRlcArqDlNack) * 100% ➢RLC Layer counters: The counters provide an indication for the total number of successful and unsuccessful transmissions at the RLC layer. pmRlcArqDlAck pmRlcArqDlNack pmRlcArqUlAck pmRlcArqUlNack ➢The following formulas can be used to determine the success rate at RLC. RLC UL success rate % = pmRlcArqUlAck/( pmRlcArqUlAck+ pmRlcArqUlNack) * 100% RLC DL success rate % = pmRlcArqDlAck/( pmRlcArqDlAck+ pmRlcArqDlNack) * 100% Figure 4-17: Exercise 2 – Answer Cont’d LZT1381950 R1A © Ericsson AB 2017 - 183 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop 4.2.5 MAC Layer counters The counters being used during HARQ transmission, are distributed per modulation scheme. ‘pmMacHarqDlAckQpsk’, ‘pmMacHarqDlNackQpsk’, ‘pmMacHarqDlDtxQpsk’, ‘pmMacHarqDlAck16qam’, ‘pmMacHarqDlNack16qam’, ‘pmMacHarqDlDtx16qam’, ‘pmMacHarqDlAck64qam’, ‘pmMacHarqDlNack64qam’, ‘pmMacHarqDlDtx64qam’, ‘pmMacHarqUlSuccQpsk’, ‘pmMacHarqUlFailQpsk’, ‘pmMacHarqUlDtxQpsk’, ‘pmMacHarqUlFail16qam’, ‘pmMacHarqUlSucc16qam’, ‘pmMacHarqUlDtx16qam’. These counters should be monitor for MAC layer retransmission per modulation scheme, Figure 4-18. › MAC Layer counters: The counters are distributed per modulation scheme being used during HARQ transmission. › pmMacHarqDlAckQpsk, pmMacHarqDlNackQpsk, pmMacHarqDlDtxQpsk › pmMacHarqDlAck16qam, pmMacHarqDlNack16qam, › pmMacHarqDlDtx16qam, pmMacHarqDlAck64qam, pmMacHarqDlNack64qam, pmMacHarqDlDtx64qam › pmMacHarqUlSuccQpsk, pmMacHarqUlFailQpsk, pmMacHarqUlDtxQpsk › pmMacHarqUlFail16qam. pmMacHarqUlSucc16qam, pmMacHarqUlDtx16qam Figure 4-18: Exercise 2 – Answer Cont’d The formula, in Figure 4-19, can be used to determine the BLER % for each type of modulation scheme. Use the respective counters for the modulation scheme that is of interest. › The following formula can be used to determine the BLER % for each type of modulation scheme. Use the respective counters for the modulation scheme that is of interest. › HARQ DL ACK Success Rate= pmMacHarqDlAck / (pmMacHarqDlAck + pmMacHarqDlNack) › HARQ UL ACK Success Rate= pmMacHarqUlSucc / (pmMacHarqUlSucc + pmMacHarqUlFail) › Example: Acknowledgement success rate for DL 16QAM. › HARQ DL ACK Success Rate 16QAM = pmMacHarqDlAck16qam/(pmMacHarqDlAck16qam+ pmMacHarqDlNack16qam) * 100% › Note: The eNodeB is designed to function at 10% BLER. The ack success rate is to be 90% or greater. Figure 4-19: Exercise 2 – Answer Cont’d - 184 - © Ericsson AB 2017 LZT1381950 R1A Issue Analysis, Improvements and Case-Studies for Integrity KPIs 4.2.6 PDCCH Congestion Counter Counter ‘pmPdcchCceUtil[0-19]’ is shown in Figure 4-20, ‘pmPdcchCceUtil’ is PDF counter of % of CCEs utilized compared with total CCEs available (at the maximum CFI permitted by ‘pdcchCfiMode’) each sub frame, considering bandwidth and antenna configuration. › PDCCH Congestion: › Counter pmPdcchCceUtil[0-19] pmPdcchCceUtil: PDF of % of CCEs utilized compared with total CCEs available (at the maximum CFI permitted by pdcchCfiMode) each sub frame, considering bandwidth and antenna configuration. [0]: utilization <= 5% [1]: 5% < utilization <= 10% [2]: 10% < utilization <= 15% [3]: 15% < utilization <= 20% [4]: 20% < utilization <= 25% [5]: 25% < utilization <= 30% [6]: 30% < utilization <= 35% [7]: 35% < utilization <= 40% [8]: 40% < utilization <= 45% [9]: 45% < utilization <= 50% [10]: 50% < utilization <= 55% [11]: 55% < utilization <= 60% [12]: 60% < utilization <= 65% [13]: 65% < utilization <= 70% [14]: 70% < utilization <= 75% [15]: 75% < utilization <= 80% [16]: 80% < utilization <= 85% [17]: 85% < utilization <= 90% [18]: 90% < utilization <= 95% [19]: 95% < utilization <= 100% Figure 4-20: Exercise 2 – Answer Cont’d 4.2.7 Uplink-Downlink PRB Utilization The Counter ‘pmPrbUtilDl[0-9],’ is a distribution that shows the downlink Physical Resource Blocks (PRB) utilization (total number of used PRB by available PRB on the Physical Downlink Shared Channel (PDSCH). ‘pmPrbUtilUl[0-9]’ is a distribution that shows the Physical Resource Blocks (PRB) utilization (total number of used PRB by available PRB) on the Physical Uplink Shared Channel (PUSCH). LZT1381950 R1A © Ericsson AB 2017 - 185 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop below Figure 4-21 illustrate the PRB utilization counter. › Downlink PRB Utilization: pmPrbUtilDl[0-9] A distribution that shows the downlink Physical Resource Blocks (PRB) utilization (total number of used PRB by available PRB on the Physical Downlink Shared Channel (PDSCH). › Uplink PRB Utilization: pmPrbUtilUl[0-9] A distribution that shows the Physical Resource Blocks (PRB) utilization (total number of used PRB by available PRB) on the Physical Uplink Shared Channel (PUSCH). PDF ranges: [0]: 0 % <= Utilization < 10 % [1]: 10 % <= Utilization < 20 % [2]: 20 % <= Utilization < 30 % [3]: 30 % <= Utilization < 40 % [4]: 40 % <= Utilization < 50 % [5]: 50 % <= Utilization < 60 % [6]: 60 % <= Utilization < 70 % [7]: 70 % <= Utilization < 80 % [8]: 80 % <= Utilization < 90 % [9]: 90 % <= Utilization Figure 4-21: Exercis 2 – Answer Cont’d 4.2.8 Counters for UEs scheduled per TTI: Internal check can be done to validate the number of users which are being scheduled per TTI, following formula can be used: ‘ulSesPerTti’=’pmSchedActivityUeUl’/’pmSchedActivityCellUl’ ‘dlSesPerTti’=’pmSchedActivityUeDl’/’pmSchedActivityCellDl’ Figure 4-22 illustrate the counters for scheduled number of UEs per TTI. › UEs scheduled per TTI: › Internal check can be done to validate the number of users which are being scheduled per TTI › The following formula can be used: – ulSesPerTti=pmSchedActivityUeUl/pmSchedActivityCellUl – dlSesPerTti=pmSchedActivityUeDl/pmSchedActivityCellDl Figure 4-22: Exercise 2 – Answer Cont’d - 186 - © Ericsson AB 2017 LZT1381950 R1A Issue Analysis, Improvements and Case-Studies for Integrity KPIs 4.3 Exercise 3: possible root cause Analyze the counters found in step 2 to determine the possible root causes for the low downlink throughput. Hint: Focus only on the cell and time line found in exercise 1 to see performance degradations. › Analyze the counters found in step 2 to determine the possible root causes for the low downlink throughput. › Hint: Focus only on the cell and time line found in exercise 1 to see performance degradations. Figure 4-23: Exercise 3: 4.3.1 Counter pmRadioRecInterferencePwr analysis-results The counter ‘pmRadioRecInterferencePwr’ has high number of pegs in bins 10 and 11 which represents the measured noise and interference is in the range of 112 to -104. Figure 4-25 and Figure 4-24 is kept blank as participants are expected to find the counter values and graph from excel sheets. And write down on blank space. › Counter pmRadioRecInterferencePwr: shows high number of pegs in bins 10 and 11 which represents the measured noise and interference is in the range of -112 to -104. Figure 4-24: Exercise 3 – Answer Cont’d LZT1381950 R1A © Ericsson AB 2017 - 187 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop 4.3.2 Counter pmRadioUeRepCQIDistr analysis-results The Counter ‘pmRadioUeRepCQIDistr’ shows the majority of the UEs using Rank Indicator 1 are reporting CQI values of 4 and 5. Figure 4-25 is kept blank participants are expected to find the counter values and graph from excel sheets. › Counter pmRadioUeRepCQIDistr shows the majority of the UEs using Rank Indicator 1 are reporting CQI values of 4 and 5. Figure 4-25: Exercise 3 – Answer Cont’d 4.3.3 Counter pmRadioUeRepCQIDistr2 analysis-results The counter ‘pmRadioUeRepCQIDistr2’ shows the majority of the UEs using Rank Indicator 2 are reporting CQI values from 2 to 5. The results are shown in Figure 4-26. › Counter pmRadioUeRepCQIDistr2 shows the majority of the UEs using Rank Indicator 2 are reporting CQI values from 2 to 5. Figure 4-26: Exercise 3 – Answer Cont’d - 188 - © Ericsson AB 2017 LZT1381950 R1A Issue Analysis, Improvements and Case-Studies for Integrity KPIs 4.3.4 PDCP layer counter analysis-results PDCP layer counter analysis shows no issues at the PDCP layer as the number of received packets ‘pmPdcpPktReceivedDl’ is equivalent to the number of transferred packets ‘pmPdcpFwdDl’. If there would have been issues, then the number of PDCP discards ‘pmPdcpDisc’ would have high counts. › PDCP layer counter analysis shows no issues at the PDCP layer as the number of received packets (pmPdcpPktReceivedDl) is equivalent to the number of transferred packets (pmPdcpFwdDl). If there would have been issues then the number of PDCP discards (pmPdcpDisc) would have high counts. Figure 4-27: Exercise 3 – Answer Cont’d LZT1381950 R1A © Ericsson AB 2017 - 189 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop 4.3.5 RLC layer counter analysis-results The success rate at RLC layer should be well above 90%. You will see from the RLC counters that in RLC UL direction there are many NACKs. If you were to compute the RLC ACK rate in both directions, you will notice that on RLC Uplink the success rate is below 90% while on downlink it’s close to 99%. This gives an indication that there is an issue on uplink. Figure 4-28. › The success rate at RLC layer should be well above 90%. You will see from the RLC counters that in RLC UL direction there are many NACKs. If you were to compute the RLC ACK rate in both directions you will notice that on RLC Uplink the success rate is below 90% while on downlink its close to 99%. This gives an indication that there is an issue on uplink. Figure 4-28: Exercise 3 – Answer Cont’d 4.3.6 Counter ‘pmMacHarq’ analysis-results for UL/DL › Counters pmMacHarq for uplink and downlink gives an indication of the modulation scheme being used. The eNodeB is designed to function at 10% BLER. From the counters we can see that most of the UEs are using QPSK modulation in downlink and uplink. You can calculate the BLER by looking at the success rate and we would like to see 90% or greater. Figure 4-29: Exercise 3 – Answer Cont’d - 190 - © Ericsson AB 2017 LZT1381950 R1A Issue Analysis, Improvements and Case-Studies for Integrity KPIs 4.3.7 MAC HARQ Counters analysis-result The blank Part in Figure 4-30 below, is an exercise for participants to find the results. › If you take a closer look at the success rate in uplink by using the raw counters you will see that in uplink for QPSK we are below the 90% target and in some of the ROPs we are below 80%. Figure 4-30: Exercise 3 – Answer Cont’d 4.3.8 Counter pmPdcchCceUtil analysis-results The analysis for the counter ‘pmPdcchCceUtil’ shows high levels (~80%) of PDCCH congestion is observed during high load., The blank Part in Figure 4-31 below, is an exercise for participants to find the results. ➢The analysis for the counter pmPdcchCceUtil shows high levels (~80%) of PDCCH congestion is observed during high load. Figure 4-31: Exercise 3 – Answer Cont’d LZT1381950 R1A © Ericsson AB 2017 - 191 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop 4.3.9 SE utilization analysis-results Number of SEs per TTI shows that when low downlink throughput occurs, the scheduler is prioritizing UEs in uplink rather than downlink. Poor UL performance using proportional fair scheduling algorithm is causing the scheduler to prioritize UEs which don`t exceed the UL average bit rate, Figure 4-32 below shows the results. › Number of SEs per TTI shows that when low downlink throughput occurs, the scheduler is prioritizing UEs in uplink rather than downlink. Poor UL performance using proportional fair scheduling algorithm is causing the scheduler to prioritize UEs which don`t exceed the UL average bit rate causing the UL to starve downlink. 3.5 3 2.5 dlSesPerTtti ulSesPerTti sumofSesPerTti 2 1.5 1 0.5 45 30 19 : 15 18 : 00 17 : 45 16 : 30 14 : 15 13 : 00 5 9: 4 12 : 0 8: 3 11 : 5 7: 1 0 Figure 4-32: Exercise 3 – Answer Cont’d 4.4 Exercise Summary With the above analysis and KPI results, it can be safely deduced that the throughput was degraded because of High Uplink Interference and Poor channel quality which leads to higher CCE aggregation causing PDCCH congestion. Scheduling algorithm set to proportional fair low causes additional problems by scheduling users on Uplink which don’t meet average bit rate throughput (lower than ‘ulMinBitRate’) so these users will be given priority by the scheduler. This causes the uplink to starve the downlink as well as UEs in poor RF are given priority over UEs in good RF. The results are discussed in Figure 4-33 below. › High Uplink Interference › Poor channel quality which leads to higher CCE aggregation causing PDCCH congestion › Scheduling algorithm set to proportional fair low causes additional problems. Figure 4-33: Exercise Summary - 192 - © Ericsson AB 2017 LZT1381950 R1A Issue Analysis, Improvements and Case-Studies for Integrity KPIs 4.5 WAY FORWARD The following recommendations, as discussed in Figure 4-34, were geared towards reducing number of retransmissions and PDCCH congestion in order to achieve higher cell downlink throughput when a high number of RRC connected users is reached in the cell. When sending an RLC PDU the UE/eNodeB may poll the eNOdeB/UE to send a STATUS report and starts the poll retransmit timer. If a timeout occurs, then the RLC PDU is retransmitted. By extending the time for polling allows additional time for the PDCCH to be used for scheduling of other UEs. This change is intended to reduce the PDCCH congestion (ie. Counter ‘pmPdcchCceUtil’) and serve other UEs. Parameters ‘tPollRetransmitUl’ and ‘tPollRetransmitDl’ Retransmissions have the highest priority within the scheduler. In highly loaded sites whereby there is high uplink interference and UEs at the cell border which are in poor cell coverage, reducing the number of RLC retransmissions for Signaling Radio Bearer and Data Radio Bearer to 8 allows UEs which require a high number of retransmissions to be dropped giving other UEs an opportunity to be scheduled. The change can cause a retainability degradation due to the reduction in the number of retransmissions. This change is intended to reduce the PDCCH congestion (ie. Counter ‘pmPdcchCceUtil’). Parameters ‘dlMaxRetxThreshold’ and ‘ulMaxRetxThreshold’ The following recommendations were geared towards reducing number of retransmissions and PDCCH congestion › By extending the time for polling allows additional time for the PDCCH to be used for scheduling of other UEs. This change is intended to reduce the PDCCH congestion (ie. Counter pmPdcchCceUtil) and serve other UEs. Parameters tPollRetransmitUl and tPollRetransmitDl › RLC retransmissions for SRB and to 8 Parameters dlMaxRetxThreshold and ulMaxRetxThreshold › PDCCH congestion (ie. Counter pmPdcchCceUtil). can be alleviated by › feature Enhanced PDCCH Link Adaptation & PDCCH Power Boost feature. › Change scheduling algorithm from proportional fair scheduling to resource fair. Figure 4-34: WAY FORWARD To reduce the CFI (Control Format Indicator) and number of CCEs (Control Channel Elements) utilization for the UE, in order to alleviate some of the PDCCH congestion. We can use few feature (ie. Counter ‘pmPdcchCceUtil’). Following features can be used ‘Enhanced PDCCH Link Adaptation’ & ‘PDCCH Power Boost feature’. PDCCH CFI mode can be defined Auto maximum. LZT1381950 R1A © Ericsson AB 2017 - 193 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop When proportional fair scheduling is used as scheduling algorithm, UEs must meet a minimum bit rate (1 Kb/s). UEs which don’t meet a minimum bit rate will be given higher priority to be scheduled which causes other UEs in good radio condition to be starved. There were UEs which wouldn’t meet the minimum bit rate (1 Kb/s) on the uplink which was causing the scheduler to prioritize these users and starving UEs with downlink data. Change scheduling algorithm from proportional fair scheduling to resource fair. Summary of Chapter 4 5 The participants should now be able to: 4 4.1 4.2 4.3 4.4 Classify the issues, analyze them for improvements and discuss few case studies for Integrity KPIs. Identify steps for UL/DL Throughput investigation & essential check Describe DL & UL throughput optimization strategy Discuss different cases of throughput & Latency degradation Analyze cases based on low throughput high latency exercises and provable solutions Figure 4-35: Summary of Chapter 4 - 194 - © Ericsson AB 2017 LZT1381950 R1A Abbreviations 5 Abbreviations 1XRTT 3GPP 3PP AAA ACIR ACK ACLR ACM ACP ACS AES AGW AIF AIR AISG AKA AM AMBR A-MPR AMR ANM ANR AP APAC API APN A-RACF ARP ARPU ARQ ARW AS A-SBG ASC ASD ASSL ASSR AuC LZT1381950 R1A operating mode of CDMA2000. 1x (the number of 1.25MHz channels) Radio Transmission Technology 3rd Generation Partnership Project 3rd Party Product Authentication, Authorization and Accounting Adjacent Channel Interference Ratio Acknowledgement Adjacent Channel Leakage Ratio Address Complete Message (ISUP Message) Automatic Cell Planning Adjacent Channel Selectivity Advanced Encryption Standard Access Gateway Auto-Integration Function Automated Integration of RBS Antenna Interface Standards Group Authentication and Key Agreement Acknowledged Mode Aggregate Maximum Bit Rate Additional Maximum Power Reduction Adaptive Multi-Rate Answer Message (ISUP Message) Automated Neighbor Relation Aggregation Proxy Asia Pacific Application Programming Interface Access Point Name Access-Resource Admission Control Function Allocation and Retention Priority Average Revenue Per User Automatic Repeat Request Add RBS Wizard Access Stratum Access SBG Antenna System Controller Automatic SW Download Adjacent Subcarrier Set Leakage Adjacent Subcarrier Set Rejection Authentication Centre © Ericsson AB 2017 - 195 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop AV B2B UA BB BCCH BCE BCH BE-DB BEM BFCP BGCF BGF BM-SC BS BSR BW C/I CA CAI CAPEX CAPEX CAS CAZAC C-BGF CCCH CCE CFR CFRA CIC CM CMC CMDB CN CO CoH COMINF Authentication Vector Back To Back User Agent Broadband Broadcast Control Channel Business Communication Enabler Broadcast Channel Back End Database Block Edge Masks Binary Floor Control Protocol Breakout Gateway Control Function Border Gateway Function Broadcast-Multicast Service Center Base Station Buffer Status Report Bandwidth Carrier-to-Interference Power Ratio Certificate Authority Customer Administration Interface Capital Expenditure Capital Expenditure Customer Administration System Constant Amplitude Zero Auto-Correlation Core Border Gateway Function Common Control Channel Control Channel Elements Counter Cipher Mode with Block Chaining Message Authentication Code Protocol Cyclic Delay Diversity Cumulative Distribution Function Call Diversion Code Division Multiple Access Carrier Ethernet The European Conference of Postal and Telecommunications Administrations Channel Feedback Report Contention Free Random Access Carrier Identification Code Configuration Management Connection Mobility Control Configuration Management Data Base Core Network Conference Owner Conference Handler Common O&M Infrastructure CO-OP Cooperative Open-OSS Project (interface also called Itf-P2P) CORBA CoSe CP CPC C-plane CPM CQI CRC Common Object Request Broker Architecture IMS Communication Services Control Plane Continous Packet Connectivity Control Plane Converged IP Messaging Channel Quality Indicator Cyclic Redundancy Check CCMP CDD CDF CDIV CDMA CE CEPT - 196 - © Ericsson AB 2017 LZT1381950 R1A Abbreviations C-RNTI Cell RNTI CS Circuit Switched CSCF Call Session Control Function CSFB Circuit Switched FallBack CSV Comma-Separated Values CTR Cell TRace CUDB Centralized User Database CW Codeword CX-AS Capability Exchange Application System DBL Dynamic Black List DBS/SABE Delay Based Scheduling/Service Aware Buffer Estimation DCCH Dedicated Control Channel DCH Dedicated Channel DCI Downlink Control Information DCN Data Communication Network DECT Digital Enhanced Cordless Telecommunications DFT Discrete Fourier Transform DFT-SDFT Spread OFDM OFDM DHCP Dynamic Host Configuration Protocol DHCP Dynamic Host Configuration Protocol Diameter represents the next generation of authentication, Diameter authorization, and accounting (AAA) controls for network access, optimized for mobile access and advanced services DL Downlink DL-SCH Downlink Shared Channel Domain Name System (Defined in STD 13, RFC 1034, RFC 1035 and DNS a number of following RFCs.) DNS Domain Name Service DRB Data Radio Bearer DRX Discontinuous Reception DSCP Differentiated Services Code Point DSL Digital Subscriber Line DTCH Dedicated Traffic Channel DTX Discontinuous Transmission DwPTS Downlink Pilot Time Slot International Public Telecommunication Numbering Plan as describe in E.164 the ITU-T Recommendation E.164. EBS Ericsson Blade System ECC Electronic Communications Committee ECGI E-UTRAN Cell Global Identifier ECM EPS Connection Management E-CSCF Emergency-CSCF E-DCH Enhanced DCH eDNS External DNS e-GEM e-GEM 2 EHPLMN EMA EMEA EMM eNB ENM LZT1381950 R1A Enhanced Generic Ericsson Magazine Enhanced Generic Ericsson Magazine 2:nd version Equivalent Home PLMN Ericsson Multi-Activation Europe, Middle East and Africa EPS Mobility Management E-UTRAN NodeB Ericsson Network Manager © Ericsson AB 2017 - 197 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop eNode B ENUM EPC EPS E-RAB ESM ETSI ETWS E-UTRA E-UTRAN EV-DO EVM EXB FCC FDD FDM FDMA FEC FFS FFT FM FMC FMX FQDN FS FTP FW GBR GCL GE GERAN GGSN GIBA GINR GMPLS GNSS GP GPRS GSM GSMA GSTN GTP GTP-C GTP-U GUI GUTI GW HA-CS HARQ HD HDVC HLR - 198 - E-UTRAN NodeB A working group within IETF that produced a standard for E.164 number and DNS. Defined in IETF RFC 2916. Ericsson Policy Control Evolved Packet System E-UTRAN Radio Access Bearer EPS Subscriber Module European Telecommunications Standards Institute Earth Quake and Tsunami Warning System Evolved UTRA Evolved UTRAN, used as synonym for LTE in the document. Evolution - Data Optimized Error Vector Magnitude Extension Blade Federal Communications Commission Frequency Division Duplex Frequency Division Multiplexing Frequency Division Multiple Access Forward Error Correction For Further Study Fast Fourier Transform Fault Management Fixed-Mobile Convergence Fault Management Expert Fully Qualified Domain Name Frame Structure File Transfer Protocol Fire Wall Guaranteed Bit Rate Generalized Chirp Like Gigabit Ethernet GSM EDGE Radio Access Network Gateway GPRS Support Node GPRS IMS Bundled Authentication Gain to Interference and Noise Ratio Generalized Multi-Protocol Label Switching Global Navigation Satellite System Guard Period General Packet Radio Service Global System for Mobile communication GSM Association General Switched Telephone Network (PSTN + PLMN) GPRS Tunneling Protocol GTP Control GTP User Data Tunneling Graphical user Interface Globally Unique Temporary Identifier Gateway High Availability Cluster Solution Hybrid ARQ High Definition High Quality Video Conferencing Home Location Register © Ericsson AB 2017 LZT1381950 R1A Abbreviations HO HOM HP HPLMN HRPD HSDPA HS-DSCH HSPA HSS HSS-FE HSUPA HTTP HTTP HTTPS HW HW IASA IBCF I-BGF ICB ICIC ICS I-CSCF ID iDNS IEEE IETF IFFT IM IMEI IMPI IMS IMSI IMSI IMS-M IMT IP IPSec IPTV Ipv4 Ipv6 IPW IRAT IRAT IS ISC ISDN ISER ISI ISIM ISM ISUP ITU LZT1381950 R1A Handover Higher Order Modulation Hewlett Packard Home PLMN High Rate Packet Data High Speed Downlink Packet Access High Speed Downlink Shared Channel High Speed Packet Access Home Subscriber Server HSS Front End High Speed Uplink Packet Access Hypertext Transfer Protocol Hypertext Transfer Protocol Hypertext Transfer Protocol over Secure Socket Layer Hardware Hardware Inter-Access Anchor Interconnection Border Control Function Interconnection Border Gateway Function Incoming Call Barring Inter-Cell Interference Coordination IMS Centralized Services Interrogating CSCF Identifier Internal DNS Institute of Electrical and Electronics Engineers Internet Engineering Task Force Inverse FFT Instant Messaging International Mobile Equipment Identity IP Multimedia Private Identity IP Multimedia subsystem International Mobile Subscriber Identity Individual Mobile Subscriber Identity IMS Messaging IMS Multimedia Telephony Internet Protocol IP Security Internet Protocol Television (Television over IP) Internet Protocol version 4. (Defined in IETF STD 5 and RFC 791). Internet Protocol version 6.(Defined in IETF RFC 2460) IPWorks Inter Radio Access Technology Inter Radio Access Technology Integrated Site IMS Service Control interface. (Defined in 3GPP TS 23.228) Integrated Services Digital Network Integrated Site Edge Router Inter Symbol Interference IMS SIM IMS Subscriber Module ISDN User Part International Telecommunications Union © Ericsson AB 2017 - 199 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop ITU-R ITU Radio Communication Sector Telecommunication Standardization Sector of the International ITU-T Telecommunications Union IWF Interworking function IWS CDMA200 InterWorking Solution JCP Java Community Process JSR Java Specification Request KPI Key Performance Indicator LAN Local Area Network LB Load Balancing LCID Logical Channel ID LCR Low Chip Rate LCR-TDD Low Chip Rate TDD LDAP Lightweight Directory Access Protocol. (Defined in IETF RFC 1777) LDC Linear Dispersion Code LDPC Low-Density Parity-check Code LED Light Emitting Diode LTE Long Term Evolution M3UA MTP Level 3 (MTP3) User Adaptation Layer MAC Medium Access Control MAP Mobile Application Part MAR/MAA Multimedia Authentication Request/Answer (Diameter Message) MBA Management Based Activation MBMS Multimedia Broadcast Multicast Service MBR Maximum Bit Rate MBSFN Multicast Broadcast Single Frequency Network MCCH Multicast Control Channel MCE Multi-cell/multicast Coordination Entity MCH Multicast Channel MCID Malicious Communication Identification MCR Malicious Communication Rejection MCS Modulation and Coding Scheme MEF Mobile Entertainment Forum MeGaCo Media Gateway Control Protocol (also referred to as H.248) MGC Media Gateway Controller MGCF Media Gateway Controller Function MGW Media Gateway MIB Management Information Base MIMO Multiple Input Multiple Output MiO Messaging in One ML-PPP Multilink point to point protocol MM Mobility Management MME Mobility Management Entity M-MGw Mobile Media Gateway MMS Multimedia Messaging Service Managed Objects interface (MOCI) MMTel Multi Media Telephony MNP Mobile Number Portability MOCI Managed Object Configuration Interface MOP Maximum Output Power MOS Mean Opinion Score MP Media Proxy MPLS Multiple Protocol Label Switching MPR Maximum Power Reduction - 200 - © Ericsson AB 2017 LZT1381950 R1A Abbreviations MRFC MRFP MRS MS MSAP MSC MSISDN MSP MSS MTAS MTCH MU-MIMO mUPE MXB NAB NACK NAI NAS NASS NAT NB NBA NCC NCL NCLI NCS NE NEM NENA NGMN NGSA NH NM NMS NMX NNI NOC NR NRT N-SBG NSP NW O&M OAM OCB OFDM OFDMA OMA OMC OOB OPEX OSS OSS-RC LZT1381950 R1A Media Resource Function Controller Media Resource Function Processor Media Resource System Management Services MCH Subframe Allocation Pattern Mobile Switching Center Mobile Subscriber Integrated Services Digital Network-Number Multi Service Proxy Mobile Softswitch Solution Multimedia Telephony Application Server Multicast Traffic Channel Multiple User-MIMO MBMS UPE Main Switching Blade Name and Address Book Negative Acknowledgement Network Access Identifier Non-Access Stratum Network Attachment Subsystem Network Address Translation Narrowband NASS Bundled Authentication Network Color Code Neighbour Cell List Node Command Line Interface Neighbouring Cell Support Network Element Network Element Manager National Emergency Number Association Next Generation Mobile Networks Next Generation Service Assurance Next Hop Key Network Management Network Management System Network level deployment of expert rules Network – Network Interface Network Operations Center Neighbor cell Relation Non Real Time Network SBG Network Server Platform Network Operation and Maintenance Operations Administration and Management Outgoing Call barring Orthogonal Frequency Division Multiplexing Orthogonal Frequency Division Multiple Access Open Mobile Alliance Operation and Maintenance Center Out of Band Operating Expenditures Operation and Support System Operation and Support System Radio and Core © Ericsson AB 2017 - 201 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop OTN P(N)CCH P2P PA PANI PAPR PAR PARC PBBTE PBC PBCH PBN PBR PC PCC PCCH PCEF PCFICH PCH PCI PCM PCO PCRF P-CSCF PDB PDCCH PDCP PDN PDN-GW PDP PDSCH PDSN PDU PGM P-GW PHICH PHR PHS PHY PLMN PM PMCH PMI PMIP PnP PoP PRACH PRB PRD P-RNTI PS PSC - 202 - Operator Terminal Network Paging (and Notification) Control Channel Peer-to-Peer Power Amplifier P-Access-Network-Info (SIP Header) Peak to Average Power Ratio Peak to Average Ratio Per Antenna Rate Control Provider Backbone Bridge Traffic Engineering Power and Battery Cabinet Physical Broadcast Channel Packet Backbone Network Prioritized Bit Rate Personal Computer Policy and Charging Control Paging Control Channel Policy Charging Enforcement Function Physical Control Format Indicator Channel Paging Channel Physical Cell ID Pulse Code Modulation Protocol Configuration Option Policy and Charging Rules Function Proxy - Call Session Control Function Packet Delay Budget Physical Downlink Control Channel Packet Data Convergence Protocol Packet Data Network Packet Data Network Gateway Packet Data Protocol Physical Downlink Shared Channel Packet Data Serving Node is a component of a CDMA2000 mobile network Protocol Data Unit Presence Group and Data Management PDN Gateway Physical Hybrid ARQ Indicator Channel Power Headroom Report Personal Handy-phone System Physical layer Public Land Mobile Network Performance Management Physical Multicast Channel Precoding Matrix Indicator Proxy Mobile IP Plug and Play Point of Presence Physical Random Access Channel Physical Resource Block Permanent Reference Document (by GSMA) Paging RNTI Packet Switched Packet Scheduling © Ericsson AB 2017 LZT1381950 R1A Abbreviations P-SCH PSD PSK PSTN PTT PUCCH PUI PUSCH QAM QCI QoS QPP QPSK RA RA RAC RACH RADIUS RAN RAN RANAP RA-RNTI RAT RB RBC RBG RBS RCS RCS-e RET RF RFC RI RLC RM RNC RNL RNTI RoHC ROP RPLMN RRC RRM RRU RS RSN RT RTCP RTP RTSP RU RX LZT1381950 R1A Primary Synchronization Channel Power Spectrum Density Pre-Shared Keys Public Switched Telephone Network Push To Talk Physical Uplink Control Channel Public User Identity Physical Uplink Shared Channel Quadrature Amplitude Modulation QoS Class Identifier Quality of Service Quadrature Permutation Polynomial Quadrature Phase Shift Keying Random Access Registration Authority Radio Admission Control Random Access Channel Remote Authentication Dial In User Service (Defined in RFC 2865, RFC 2866 and RFC 2869) Radio Access Network Radio Access Network RAN Application Part Random Access RNTI Radio Access Technology Radio Bearer Radio Bearer Control Radio Bearer Group Radio Base Station Rich Communication Suite Rich Communication Suite-enhanced Remote Electrical Tilt Radio Frequency Request for Comment Rank Indicator Radio Link Control Rate Matching Radio Network Controller Radio Network Layer Radio Network Temporary Identifier Robust Header Compression Recording Output Periods Registered PLMN Radio Resource Control Radio Resource Management Radio Remote Unit Reference Signal Retransmission SN Real Time RTP Control Protocol Real Time Protocol Real Time Streaming Protocol Resource Unit Receiver © Ericsson AB 2017 - 203 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop S1-MME S1-U SAE SAP SAPC SAR/SAA SB SBC SBG SCC-AS SCCH SCCP SCEP SC-FDMA SCH S-CSCF SCTP SDES SDF SDH SDMA SDP SDU SeGW SEM SFN SFP SFTP S-FTP SGC SGSN S-GW SI SIB SID SIM SINR SIP SI-RNTI SIS SISO SLA SLF SLO SM SMO SMRS SMS SMS SN SNF SNR SOAP - 204 - S1 for the control plane S1 for the user plane System Architecture Evolution Service Access Point Service Aware Policy Controller Server Assignment Request/Answer (Diameter Message) Scheduling Block Session Border Controller Session Border Gateway Service Centralization and Continuity Application Server Shared Control Channel Signaling Connection Control Part Simple Certificate Enrolment Protocol Single Carrier – Frequency Division Multiple Access Synchronization Channel Serving CSCF Stream Control Transmission Protocol Session Description Protocol Security Descriptions Service Data Flow Synchronous Digital Hierarchy Spatial Division Multiple Access Session Description Protocol Service Data Unit Security Gateway Spectrum Emission Mask System Frame Number Small Form factor Pluggable Secure File Transfer Protocol Secure File transfer protocol Session Gateway Controller Serving GPRS Support Node Serving Gateway System Information System Information Block Silence Insertion Descriptor Subscriber Identity Module Signal to Interference and Noise Ratio Session Initiation Protocol System Info RNTI Site Infrastructure Support Single Input Single Output Service Level Agreement Subscriber Location Function Service Level Objectives Session Management Software Manager Organizer Software Management Repository Short Message Service Short Message Service Sequence Number Service Network Framework Signal to Noise Ratio Simple Object Access Protocol © Ericsson AB 2017 LZT1381950 R1A Abbreviations SON SOX S-PARC SPDF SPID SQL SR SRB SRTP SRVCC SS7 S-SCH SSH SSL SSLIOP SSO SSP SU SU-MIMO SW SW TA T-ADS TAS TAU TB TBD TCP TDD TDM TF TFCI TFP TFT TISPAN TLA TLP TLS TM TMA TMO TNL TPC TSP TTI TTM TX UA UAC UCI UE UETR LZT1381950 R1A Self-Organizing Networks Simple Outline XML Selective PARC Session Policy Decision Function Subscriber Profile ID for RAT/Frequency Priority Structured Query Language Scheduling Request Signaling Radio Bearer Secure Real-Time Protocol Single Radio Voice Call Continuity Signaling System, No 7 Secondary Synchronization Channel Secure Shell Secure Sockets Layer IIOP over SSL Single Sign On Self-Service Portal Scheduling Unit Single-User MIMO Software Soft Ware Tracking Area Terminating-Access Domain Selection Telephony Application Server Tracking Area Update Transport Block To Be Decided Transmission Control Protocol Time Division Duplex Time-Division Multiplexing Transport Format Transport Format Combination Indicator Traffic Forwarding Policy Traffic Flow Template Telecoms & Internet converged Services & Protocols for Advanced Networks Three Letter Acronym TEMS LinkPlanner Transport Layer Security Transparent Mode Tower Mounted Amplifier T-Mobile International AG Transport Network Layer Transmit Power Control Ericsson Telecom Server Platform Transmission Time Interval Time To Market Transmitter User Agent. An endpoint in a SIP based network. User Agent Client Uplink Control Information User Equipment UE Trace © Ericsson AB 2017 - 205 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop UL UL-SCH UM UMTS UNI UP UPE U-plane UpPTS URA URI URL USIM UTRA UTRAN VAD VoHSPA VoIP VoLTE VPLMN VRB WAP WAPECS WCDMA WCG WDM WiMAX WUIGM X2-C X2-U XCAP XDMS XML XML - 206 - Uplink Uplink Shared Channel Unacknowledged Mode Universal Mobile Telecommunication System User-to-Network Interface User Plane User Plane Entity User plane Uplink Pilot Time Slot UTRAN Routing Area Uniform Resource Identifier. (Defined in RFC 2396) Uniform Resource Locator. The address of a file (resource) accessible on the Internet. (Defined in RFC 2396) Universal SIM (3G SIM) UMTS Terrestrial Radio Access UMTS Terrestrial Radio Access Network Voice Activity Detector Voice over HSPA Voice over IP Voice over LTE Visited PLMN Virtual Resource Block Wireless Access Protocol Wireless Access Policy for Electronic Communications Services Wideband Code Division Multiple Access Web Communication Gateway Wavelength Division Multiplexing Worldwide Interoperability for Microwave Access Web User Interface for Group and Data Management X2-Control plane X2-User plane XML Configuration Access Protocol XML Document Management Server extensible Markup Language Extensible Markup Language © Ericsson AB 2017 LZT1381950 R1A Index 6 Index 3rd Generation Partnership Project, 12, 13, 14, 16, 36, 140, 141, 145, 148, 151, 154, 156, 158, 166, 178, 195, 199 3rd Party Product, 195 A working group within IETF that produced a standard for E.164 number and DNS. Defined in IETF RFC 2916., 198 Access Gateway, 195 Access Point Name, 139, 195 Access SBG, 195 Access Stratum, 195 Access-Resource Admission Control Function, 195 Acknowledged Mode, 118, 195 Acknowledgement, 154, 190, 195 Adaptive Multi-Rate, 195 Add RBS Wizard, 195 Additional Maximum Power Reduction, 195 Address Complete Message (ISUP Message), 195 Adjacent Channel Interference Ratio, 195 Adjacent Channel Leakage Ratio, 195 Adjacent Channel Selectivity, 195 Adjacent Subcarrier Set Leakage, 195 Adjacent Subcarrier Set Rejection, 195 Advanced Encryption Standard, 195 Aggregate Maximum Bit Rate, 139, 195 Aggregation Proxy, 195 Allocation and Retention Priority, 195 Answer Message (ISUP Message), 195 Antenna Interface Standards Group, 195 Antenna System Controller, 195 Application Programming Interface, 195 Asia Pacific, 195 Authentication and Key Agreement, 195 Authentication Centre, 195 Authentication Vector, 196 LZT1381950 R1A Authentication, Authorization and Accounting, 195, 197 Auto-Integration Function, 195 Automated Integration of RBS, 195 Automated Neighbor Relation, 29, 53, 58, 59, 89, 91, 92, 93, 94, 95, 97, 98, 99, 100, 101, 102, 103, 176, 195 Automatic Cell Planning, 195 Automatic Repeat Request, 195, 198, 202 Automatic SW Download, 195 Average Revenue Per User, 195 Back End Database, 196 Back To Back User Agent, 196 Bandwidth, 153, 196 Base Station, 196 Binary Floor Control Protocol, 196 Block Edge Masks, 196 Border Gateway Function, 196 Breakout Gateway Control Function, 196 Broadband, 168, 169, 196 Broadcast Channel, 196 Broadcast Control Channel, 173, 196 Broadcast-Multicast Service Center, 196 Buffer Status Report, 141, 196 Business Communication Enabler, 196 C/I, 143 Call Diversion, 196 Call Session Control Function, 197, 199, 204 Capability Exchange Application System, 197 Capital Expenditure, 196 Carrier Ethernet, 196 Carrier Identification Code, 196 Carrier-to-Interference Power Ratio, 143, 196 CDMA200 InterWorking Solution, 200 Cell RNTI, 197 Cell TRace, 197 Centralized User Database, 197 © Ericsson AB 2017 - 207 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop Certificate Authority, 18, 133, 145, 153, 157, 158, 196 Channel Feedback Report, 196 Channel Quality Indicator, 127, 137, 145, 154, 188, 196 Circuit Switched, 103, 111, 197 Circuit Switched FallBack, 103, 104, 105, 106, 107, 108, 109, 110, 197 Code Division Multiple Access, 26, 52, 196 Codeword, 197 Comma-Separated Values, 197 Common Control Channel, 196 Common O&M Infrastructure, 196 Common Object Request Broker Architecture, 196 Conference Handler, 196 Conference Owner, 196 Configuration Management, 91, 196 Configuration Management Data Base, 196 Connection Mobility Control, 196 Constant Amplitude Zero Auto-Correlation, 196 Contention Free Random Access, 37, 196 Continous Packet Connectivity, 196 Control Channel Elements, 192, 196 Control Plane, 196 Converged IP Messaging, 196 Cooperative Open-OSS Project (interface also called Itf-P2P), 196 Core Border Gateway Function, 196 Core Network, 105, 106, 108, 196 Counter Cipher Mode with Block Chaining Message Authentication Code Protocol, 196 CRS, 166 Cumulative Distribution Function, 196 Customer Administration Interface, 196 Customer Administration System, 196 Cyclic Delay Diversity, 196 Cyclic Redundancy Check, 196 Data Communication Network, 197 Data Radio Bearer, 32, 117, 118, 119, 120, 121, 124, 197 Dedicated Channel, 197 Dedicated Control Channel, 197 Dedicated Traffic Channel, 197 Delay Based Scheduling/Service Aware Buffer Estimation, 197 DFT Spread OFDM, 197 Diameter represents the next generation of authentication, authorization, and accounting (AAA) controls for network - 208 - access, optimized for mobile access and advanced services, 197, 200, 204 Differentiated Services Code Point, 140, 197 Digital Enhanced Cordless Telecommunications, 197 Digital Subscriber Line, 197 Discontinuous Reception, 92, 101, 102, 152, 175, 176, 177, 178, 197 Discontinuous Transmission, 197 Discrete Fourier Transform, 197 Domain Name Service, 197, 198, 199 Domain Name System (Defined in STD 13, RFC 1034, RFC 1035 and a number of following RFCs.), 197, 198, 199 Downlink, 18, 29, 36, 49, 87, 101, 117, 118, 119, 127, 128, 129, 131, 132, 133, 137, 138, 140, 144, 145, 146, 147, 151, 153, 154, 155, 156, 157, 158, 161, 165, 168, 172, 177, 181, 182, 183, 190, 197 Downlink Control Information, 197 Downlink Pilot Time Slot, 197 Downlink Shared Channel, 197 Dynamic Black List, 197 Dynamic Host Configuration Protocol, 197 Earth Quake and Tsunami Warning System, 198 Electronic Communications Committee, 197 Emergency-CSCF, 197 Enhanced DCH, 197 Enhanced Generic Ericsson Magazine, 197 Enhanced Generic Ericsson Magazine 2 nd version, 197 EPS Connection Management, 197 EPS Mobility Management, 197 EPS Subscriber Module, 198 Equivalent Home PLMN, 197 Ericsson Blade System, 197 Ericsson Multi-Activation, 197 Ericsson Network Manager, 197 Ericsson Policy Control, 198 Ericsson Telecom Server Platform, 205 Error Vector Magnitude, 145, 146, 198 Europe, Middle East and Africa, 197 European Telecommunications Standards Institute, 198 E-UTRAN Cell Global Identifier, 100, 101, 197 E-UTRAN NodeB, 33, 65, 90, 96, 110, 150, 158, 159, 160, 161, 176, 197, 198 E-UTRAN Radio Access Bearer, 198 Evolution - Data Optimized, 198 Evolved Packet System, 103, 139, 197, 198 © Ericsson AB 2017 LZT1381950 R1A Index Evolved UTRA, 12, 13, 14, 15, 16, 23, 100, 103, 197, 198 Evolved UTRAN, used as synonym for LTE in the document., 12, 16, 23, 100, 103, 197, 198 extensible Markup Language, 205, 206 Extensible Markup Language, 205, 206 Extension Blade, 198 External DNS, 197 Fast Fourier Transform, 198, 199 Fault Management, 198 Fault Management Expert, 198 Federal Communications Commission, 198 File Transfer Protocol, 198 Fire Wall, 198 Fixed-Mobile Convergence, 198 For Further Study, 198 Forward Error Correction, 198 Frame Structure, 198 Frequency Division Duplex, 23, 59, 94, 149, 153, 156, 159, 165, 198 Frequency Division Multiple Access, 198 Frequency Division Multiplexing, 198 Fully Qualified Domain Name, 198 Gain to Interference and Noise Ratio, 198 Gateway, 198 Gateway GPRS Support Node, 198 General Packet Radio Service, 198, 204 General Switched Telephone Network (PSTN + PLMN), 198 Generalized Chirp Like, 198 Generalized Multi-Protocol Label Switching, 198 Gigabit Ethernet, 198 Global Navigation Satellite System, 198 Global System for Mobile communication, 91, 92, 198 Globally Unique Temporary Identifier, 198 GPRS IMS Bundled Authentication, 198 GPRS Tunneling Protocol, 198 Graphical user Interface, 198 GSM Association, 198, 202 GSM EDGE Radio Access Network, 26, 52, 59, 103, 198 GTP Control, 198 GTP User Data Tunneling, 198 Guaranteed Bit Rate, 168, 198 Guard Period, 198 Handover, 29, 30, 31, 32, 33, 34, 35, 38, 39, 40, 41, 42, 45, 46, 52, 55, 64, 65, 66, 68, LZT1381950 R1A 69, 70, 71, 72, 73, 75, 79, 83, 87, 89, 90, 91, 92, 93, 95, 166, 199 Hardware, 199 Hewlett Packard, 199 High Availability Cluster Solution, 198 High Definition, 198 High Quality Video Conferencing, 198 High Rate Packet Data, 199 High Speed Downlink Packet Access, 199 High Speed Downlink Shared Channel, 199 High Speed Packet Access, 199, 206 High Speed Uplink Packet Access, 199 Higher Order Modulation, 138, 147, 148, 199 Home Location Register, 198 Home PLMN, 199 Home Subscriber Server, 199 HSS Front End, 199 Hybrid ARQ, 154, 157, 160, 184, 191, 198 Hypertext Transfer Protocol, 199 Hypertext Transfer Protocol over Secure Socket Layer, 199 Identifier, 23, 199, 200, 202, 205 IIOP over SSL, 205 IMS Centralized Services, 199 IMS Communication Services, 196 IMS Messaging, 199 IMS Multimedia Telephony, 199 IMS Service Control interface. (Defined in 3GPP TS 23.228), 199 IMS SIM, 199 IMS Subscriber Module, 199 Incoming Call Barring, 199 Individual Mobile Subscriber Identity, 199 Instant Messaging, 199 Institute of Electrical and Electronics Engineers, 199 Integrated Services Digital Network, 199 Integrated Site, 199 Integrated Site Edge Router, 199 Inter Radio Access Technology, 14, 25, 26, 27, 28, 40, 41, 42, 52, 53, 55, 68, 69, 70, 78, 79, 80, 81, 82, 83, 93, 95, 199 Inter Symbol Interference, 199 Inter-Access Anchor, 199 Inter-Cell Interference Coordination, 199 Interconnection Border Control Function, 199 Interconnection Border Gateway Function, 199 Internal DNS, 199 International Mobile Equipment Identity, 199 International Mobile Subscriber Identity, 199 © Ericsson AB 2017 - 209 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop International Public Telecommunication Numbering Plan as describe in the ITU-T Recommendation E.164., 197, 198 International Telecommunications Union, 197, 199, 200 Internet Engineering Task Force, 198, 199, 200 Internet Protocol, 33, 168, 196, 199, 202, 206 Internet Protocol Television (Television over IP), 199 Internet Protocol version 4. (Defined in IETF STD 5 and RFC 791)., 199 Internet Protocol version 6.(Defined in IETF RFC 2460), 199 Interrogating CSCF, 199 Interworking function, 200 Inverse FFT, 199 IP Multimedia Private Identity, 199 IP Multimedia subsystem, 196, 198, 199 IP Security, 199 IPWorks, 199 ISDN User Part, 195, 199 ITU Radio Communication Sector, 200 Java Community Process, 200 Java Specification Request, 200 Key Performance Indicator, 12, 25, 28, 44, 50, 54, 68, 72, 84, 90, 99, 105, 111, 116, 119, 121, 123, 126, 127, 128, 151, 172, 179, 192, 200 Light Emitting Diode, 200 Lightweight Directory Access Protocol. (Defined in IETF RFC 1777), 200 Linear Dispersion Code, 200 Load Balancing, 200 Local Area Network, 200 Logical Channel ID, 200 Long Term Evolution, 11, 12, 18, 23, 25, 26, 28, 33, 35, 36, 38, 40, 44, 59, 60, 68, 69, 70, 79, 80, 89, 91, 92, 93, 99, 105, 109, 110, 111, 112, 115, 121, 127, 140, 141, 144, 147, 148, 151, 152, 154, 161, 165, 167, 174, 198, 200, 206 Low Chip Rate, 200 Low Chip Rate TDD, 200 Low-Density Parity-check Code, 200 Main Switching Blade, 201 Malicious Communication Identification, 200 Malicious Communication Rejection, 200 Managed Object Configuration Interface, 200 Management Based Activation, 200 Management Information Base, 200 Management Services, 201 - 210 - Maximum Bit Rate, 200 Maximum Output Power, 200 Maximum Power Reduction, 200 MBMS UPE, 201 MCH Subframe Allocation Pattern, 201 Mean Opinion Score, 200 Media Gateway, 200 Media Gateway Control Protocol (also referred to as H.248), 200 Media Gateway Controller, 200 Media Gateway Controller Function, 200 Media Proxy, 30, 31, 33, 200 Media Resource Function Controller, 201 Media Resource Function Processor, 201 Media Resource System, 201 Medium Access Control, 136, 145, 157, 184, 191, 200 Messaging in One, 200 MIMO, 166 Mobile Application Part, 200 Mobile Entertainment Forum, 200 Mobile Media Gateway, 200 Mobile Number Portability, 200 Mobile Softswitch Solution, 201 Mobile Subscriber Integrated Services Digital Network-Number, 201 Mobile Switching Center, 201 Mobility Management, 200 Mobility Management Entity, 33, 40, 41, 93, 200 Modulation and Coding Scheme, 148, 200 MTP Level 3 (MTP3) User Adaptation Layer, 200 Multi Media Telephony, 200 Multi Service Proxy, 201 Multicast Broadcast Single Frequency Network, 200 Multicast Channel, 200, 201 Multicast Control Channel, 200 Multicast Traffic Channel, 201 Multi-cell/multicast Coordination Entity, 200 Multilink point to point protocol, 200 Multimedia Authentication Request/Answer (Diameter Message), 200 Multimedia Broadcast Multicast Service, 200, 201 Multimedia Messaging Service Managed Objects interface (MOCI), 200 Multimedia Telephony Application Server, 201 © Ericsson AB 2017 LZT1381950 R1A Index Multiple Input Multiple Output, 131, 133, 138, 145, 147, 154, 165, 166, 167, 182, 200, 201, 205 Multiple Protocol Label Switching, 200 Multiple User-MIMO, 148, 150, 201 Name and Address Book, 201 Narrowband, 201 NASS Bundled Authentication, 201 National Emergency Number Association, 201 Negative Acknowledgement, 136, 154, 201 Neighbor cell Relation, 95, 201 Neighbour Cell List, 201 Neighbouring Cell Support, 201 Network, 201 Network – Network Interface, 201 Network Access Identifier, 201 Network Address Translation, 201 Network Attachment Subsystem, 201 Network Color Code, 201 Network Element, 201 Network Element Manager, 201 Network level deployment of expert rules, 201 Network Management, 201 Network Management System, 201 Network Operations Center, 201 Network SBG, 201 Network Server Platform, 201 Next Generation Mobile Networks, 201 Next Generation Service Assurance, 201 Next Hop Key, 201 Node Command Line Interface, 201 Non Real Time, 201 Non-Access Stratum, 201 Open Mobile Alliance, 201 Operating Expenditures, 201 operating mode of CDMA2000. 1x (the number of 1.25MHz channels) Radio Transmission Technology, 195 Operation and Maintenance, 196, 201 Operation and Maintenance Center, 201 Operation and Support System, 91, 98, 110, 141, 196, 201 Operation and Support System Radio and Core, 91, 141, 201 Operations Administration and Management, 201 Operator Terminal Network, 202 Orthogonal Frequency Division Multiple Access, 201 LZT1381950 R1A Orthogonal Frequency Division Multiplexing, 151, 197, 201 Out of Band, 201 Outgoing Call barring, 201 P-Access-Network-Info (SIP Header), 202 Packet Backbone Network, 202 Packet Data Convergence Protocol, 85, 117, 118, 123, 124, 165, 183, 189, 202 Packet Data Network, 202 Packet Data Network Gateway, 202 Packet Data Protocol, 202 Packet Data Serving Node is a component of a CDMA2000 mobile network, 202 Packet Delay Budget, 168, 202 Packet Scheduling, 202 Packet Switched, 168, 202 Paging (and Notification) Control Channel, 202 Paging Channel, 202 Paging Control Channel, 202 Paging RNTI, 202 PDN Gateway, 202 Peak to Average Power Ratio, 202 Peak to Average Ratio, 202 Peer-to-Peer, 196, 202 Per Antenna Rate Control, 202, 205 Performance Management, 165, 180, 202 Permanent Reference Document (by GSMA), 202 Personal Computer, 202 Personal Handy-phone System, 202 Physical Broadcast Channel, 202 Physical Cell ID, 29, 35, 37, 53, 65, 89, 90, 91, 92, 93, 95, 99, 100, 101, 102, 202 Physical Control Format Indicator Channel, 202 Physical Downlink Control Channel, 101, 127, 133, 152, 153, 166, 168, 177, 185, 191, 192, 193, 202 Physical Downlink Shared Channel, 87, 132, 166, 185, 202 Physical Hybrid ARQ Indicator Channel, 202 Physical layer, 202 Physical Multicast Channel, 202 Physical Random Access Channel, 202 Physical Resource Block, 127, 132, 133, 134, 151, 152, 154, 156, 185, 186, 202 Physical Uplink Control Channel, 88, 135, 153, 156, 157, 173, 203 © Ericsson AB 2017 - 211 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop Physical Uplink Shared Channel, 36, 132, 135, 148, 150, 151, 154, 162, 173, 182, 185, 203 Plug and Play, 202 Point of Presence, 202 Policy and Charging Control, 202 Policy and Charging Rules Function, 202 Policy Charging Enforcement Function, 202 Power Amplifier, 202 Power and Battery Cabinet, 202 Power Headroom Report, 202 Power Spectrum Density, 174, 203 Precoding Matrix Indicator, 202 Presence Group and Data Management, 202 Pre-Shared Keys, 203 Primary Synchronization Channel, 203 Prioritized Bit Rate, 202 Protocol Configuration Option, 202 Protocol Data Unit, 136, 193, 202 Provider Backbone Bridge Traffic Engineering, 202 Proxy - Call Session Control Function, 202 Proxy Mobile IP, 202 Public Land Mobile Network, 33, 93, 197, 198, 199, 202, 203, 206 Public Switched Telephone Network, 198, 203 Public User Identity, 203 Pulse Code Modulation, 202 Push To Talk, 203 QoS, 143 QoS Class Identifier, 23, 42, 53, 59, 60, 79, 116, 121, 124, 140, 141, 168, 169, 174, 175, 176, 177, 203 Quadrature Amplitude Modulation, 131, 133, 138, 145, 146, 147, 148, 149, 150, 151, 154, 156, 203 Quadrature Permutation Polynomial, 203 Quadrature Phase Shift Keying, 138, 203 Quality of Service, 60, 140, 141, 142, 143, 176, 203 Radio Access Network, 121, 140, 142, 160, 165, 203 Radio Access Technology, 14, 15, 23, 94, 139, 203, 205 Radio Admission Control, 203 Radio Base Station, 40, 41, 96, 116, 118, 119, 120, 121, 122, 123, 124, 126, 140, 141, 142, 168, 195, 203 Radio Bearer, 203 Radio Bearer Control, 203 Radio Bearer Group, 203 - 212 - Radio Frequency, 29, 75, 166, 181, 192, 203 Radio Link Control, 116, 118, 128, 136, 183, 190, 193, 203 Radio Network Controller, 40, 203 Radio Network Layer, 203 Radio Network Temporary Identifier, 197, 202, 203, 204 Radio Remote Unit, 29, 33, 203 Radio Resource Control, 26, 40, 45, 46, 64, 65, 66, 67, 87, 88, 90, 91, 93, 107, 108, 157, 178, 193, 203 Radio Resource Management, 23, 203 RAN Application Part, 203 Random Access, 37, 174, 203 Random Access Channel, 37, 38, 203 Random Access RNTI, 37, 203 Rank Indicator, 127, 203 Rate Matching, 203 Real Time, 203 Real Time Protocol, 203 Real Time Streaming Protocol, 203 Receiver, 150, 203 Recording Output Periods, 119, 121, 122, 126, 179, 180, 203 Reference Signal, 203 Registered PLMN, 203 Registration Authority, 37, 174, 203 Remote Authentication Dial In User Service (Defined in RFC 2865, RFC 2866 and RFC 2869), 203 Remote Electrical Tilt, 203 Request for Comment, 197, 198, 199, 200, 203, 206 Resource Unit, 203 Retransmission SN, 203 Rich Communication Suite, 203 Rich Communication Suite-enhanced, 203 Robust Header Compression, 203 RTP Control Protocol, 203 S1 for the control plane, 204 S1 for the user plane, 204 Scheduling Block, 204 Scheduling Request, 88, 205 Scheduling Unit, 205 Secondary Synchronization Channel, 205 Secure File transfer protocol, 204 Secure File Transfer Protocol, 204 Secure Real-Time Protocol, 205 Secure Shell, 205 Secure Sockets Layer, 205 Security Gateway, 204 Selective PARC, 205 © Ericsson AB 2017 LZT1381950 R1A Index Self-Organizing Networks, 99, 205 Self-Service Portal, 205 Sequence Number, 203, 204 Server Assignment Request/Answer (Diameter Message), 204 Service Access Point, 204 Service Aware Policy Controller, 204 Service Centralization and Continuity Application Server, 204 Service Data Flow, 204 Service Data Unit, 118, 204 Service Level Agreement, 204 Service Level Objectives, 204 Service Network Framework, 204 Serving CSCF, 204 Serving Gateway, 93, 140, 204 Serving GPRS Support Node, 40, 204 Session Border Controller, 204 Session Border Gateway, 195, 201, 204 Session Description Protocol, 204 Session Description Protocol Security Descriptions, 204 Session Gateway Controller, 204 Session Initiation Protocol, 202, 204, 205 Session Management, 204 Session Policy Decision Function, 205 Shared Control Channel, 204 Short Message Service, 204 Signal to Interference and Noise Ratio, 36, 127, 128, 135, 148, 150, 158, 204 Signal to Noise Ratio, 204 Signaling Connection Control Part, 204 Signaling Radio Bearer, 32, 205 Signaling System, No 7, 205 Silence Insertion Descriptor, 204 Simple Certificate Enrolment Protocol, 204 Simple Object Access Protocol, 204 Simple Outline XML, 205 Single Carrier – Frequency Division Multiple Access, 204 Single Input Single Output, 204 Single Radio Voice Call Continuity, 205 Single Sign On, 205 Single-User MIMO, 205 Site Infrastructure Support, 204 Small Form factor Pluggable, 204 Soft Ware, 33, 195, 205 Software, 33, 195, 205 Software Management Repository, 204 Software Manager Organizer, 204 Spatial Division Multiple Access, 204 LZT1381950 R1A Spectrum Emission Mask, 204 Stream Control Transmission Protocol, 93, 204 Structured Query Language, 205 Subscriber Identity Module, 199, 204, 206 Subscriber Location Function, 204 Subscriber Profile ID for RAT/Frequency Priority, 23, 205 Synchronization Channel, 204 Synchronous Digital Hierarchy, 204 System Architecture Evolution, 204 System Frame Number, 204 System Info RNTI, 204 System Information, 204 System Information Block, 106, 107, 152, 204 Telecommunication Standardization Sector of the International Telecommunications Union, 197, 200 Telecoms & Internet converged Services & Protocols for Advanced Networks, 205 Telephony Application Server, 205 TEMS LinkPlanner, 205 Terminating-Access Domain Selection, 205 The European Conference of Postal and Telecommunications Administrations, 196 Three Letter Acronym, 205 Time Division Duplex, 23, 59, 149, 156, 165, 200, 205 Time To Market, 205 Time-Division Multiplexing, 205 T-Mobile International AG, 205 To Be Decided, 205 Tower Mounted Amplifier, 205 Tracking Area, 205 Tracking Area Update, 205 Traffic Flow Template, 205 Traffic Forwarding Policy, 205 Transmission Control Protocol, 205 Transmission Time Interval, 118, 120, 154, 156, 157, 160, 186, 192, 205 Transmit Power Control, 205 Transmitter, 182, 205 Transparent Mode, 205 Transport Block, 205 Transport Format, 205 Transport Format Combination Indicator, 205 Transport Layer Security, 205 Transport Network Layer, 205 UE Trace, 205 UMTS Terrestrial Radio Access, 198, 206 © Ericsson AB 2017 - 213 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop UMTS Terrestrial Radio Access Network, 12, 16, 59, 103, 108, 109, 110, 198, 206 Unacknowledged Mode, 118, 206 Uniform Resource Identifier. (Defined in RFC 2396), 206 Uniform Resource Locator. The address of a file (resource) accessible on the Internet. (Defined in RFC 2396), 206 Universal Mobile Telecommunication System, 69, 78, 79, 206 Universal SIM (3G SIM), 206 Uplink, 18, 29, 36, 101, 120, 128, 129, 130, 131, 132, 133, 135, 136, 137, 139, 140, 144, 148, 149, 150, 151, 153, 154, 156, 157, 158, 162, 163, 168, 173, 174, 181, 182, 183, 190, 192, 206 Uplink Control Information, 205 Uplink Pilot Time Slot, 206 Uplink Shared Channel, 206 User Agent Client, 205 User Agent. An endpoint in a SIP based network., 205 User Equipment, 12, 14, 15, 16, 21, 23, 24, 25, 26, 28, 29, 31, 32, 39, 40, 41, 45, 46, 52, 53, 55, 56, 59, 60, 65, 67, 68, 69, 71, 79, 91, 92, 93, 94, 95, 100, 101, 102, 103, 106, 107, 109, 110, 111, 112, 116, 117, 118, 119, 120, 121, 128, 136, 139, 140, 141, 142, 145,147, 149, 151, 152, 153, - 214 - 154, 157, 158, 159, 160, 162, 165, 166, 173, 175, 176, 177, 178, 193, 205 User plane, 206 User Plane, 206 User Plane Entity, 201, 206 User-to-Network Interface, 206 UTRAN Routing Area, 206 Virtual Resource Block, 206 Visited PLMN, 206 Voice Activity Detector, 206 Voice over HSPA, 206 Voice over IP, 168, 206 Voice over LTE, 153, 206 Wavelength Division Multiplexing, 206 Web Communication Gateway, 206 Web User Interface for Group and Data Management, 206 Wideband Code Division Multiple Access, 18, 25, 26, 40, 52, 59, 78, 79, 91, 92, 106, 107, 110, 111, 206 Wireless Access Policy for Electronic Communications Services, 206 Wireless Access Protocol, 206 Worldwide Interoperability for Microwave Access, 206 X2-Control plane, 206 X2-User plane, 206 XML Configuration Access Protocol, 206 XML Document Management Server, 206 © Ericsson AB 2017 LZT1381950 R1A Table of Figure 7 Table of Figures Figure 1-1: Objective of Chapter 1 ................................................................................................ 11 Figure 1-2: Review: Cell Selection (S-Criterion) ............................................................................ 13 Figure 1-3: Cell Reselection (R-Criteria) ....................................................................................... 13 Figure 1-4: Cell Reselection Evaluation Process ........................................................................... 14 Figure 1-5: Priority Based Cell Reselection ................................................................................... 15 Figure 1-6: Speed Dependent Scaling of Cell Reselection ............................................................ 17 Figure 1-7: Cell selection/Reselection parameters ........................................................................ 17 Figure 1-8: Case-1: Cell Reselection Parameter Example ........................................................... 18 Figure 1-9: Parameter example Case-2 ........................................................................................ 19 Figure 1-10: Case 2: 3CC idle mode ............................................................................................. 19 Figure 1-11: Case 3: Initial Traffic Balance Issue & resolution on MC site ..................................... 20 Figure 1-12: Case 4: Key finding of improper parameter setting at one operator ........................... 21 Figure 1-13: Case 4: ‘qRxlevmin’ Setting ...................................................................................... 21 Figure 1-14: Case 4: Sintrasearch setting ..................................................................................... 22 Figure 1-15: Case 4: Snonintrasearch setting ............................................................................... 22 Figure 1-16: Case 4: Summary of change ..................................................................................... 23 Figure 1-17: Idle mode behavior with IFLB functionality ................................................................ 24 Figure 1-18: Idle mode with Sticky carrier methodology ................................................................ 24 Figure 1-19: EUTRAN Mobility KPI ............................................................................................... 25 Figure 1-20: Review: LTE Events .................................................................................................. 26 Figure 1-21: Review: Coverage triggered mobility ......................................................................... 27 Figure 1-22: Review: “Basic” Mobility Control in Poor Coverage Operation ................................... 28 Figure 1-23: Mobility Issue Analysis .............................................................................................. 29 Figure 1-24: Mobility issue analysis HO Prep & Exec Failures ..................................................... 30 Figure 1-25: Intra/Inter Handover Prep fail issue: possible cause.................................................. 30 Figure 1-26: Intra/Inter Handover Prep fail issue: possible cause.................................................. 31 Figure 1-27: Intra/Inter Handover Prep fail issue: possible cause.................................................. 31 Figure 1-28: Intra/Inter Handover Prep fail issue: possible cause.................................................. 32 Figure 1-29: Intra/Inter Handover Prep fail issue: possible cause.................................................. 34 Figure 1-30: Intra/Inter Handover Prep fail issue: possible cause.................................................. 34 Figure 1-31: LTE Intra/Inter Handover Execution Failure: possible cause ..................................... 35 Figure 1-32: LTE Intra/Inter Handover Execution Failure: possible cause ..................................... 35 Figure 1-33: LTE Intra/Inter Handover Execution Failure: possible cause ..................................... 36 Figure 1-34: LTE Intra/Inter Handover Execution Failure: possible cause ..................................... 36 Figure 1-35: Uplink Interference .................................................................................................... 37 Figure 1-36: LTE Intra/Inter Handover Execution Failure: possible cause ..................................... 38 Figure 1-37: HO Exec Fail: Contention Based Random Access Success Rate ............................. 38 Figure 1-38: HO Exec Fail: Overshooting Cell ............................................................................... 39 LZT1381950 R1A © Ericsson AB 2017 - 215 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop Figure 1-39: Intra/Inter Frequency HO Optimization ...................................................................... 39 Figure 1-40: IRAT Handover HO optimization ............................................................................... 41 Figure 1-41: IRAT/Inter Frequency HO optimization ...................................................................... 42 Figure 1-42: IRAT/Inter Frequency Session continuity optimization .............................................. 42 Figure 1-43: Inter-frequency load balancing (IFLB) ....................................................................... 43 Figure 1-44: LTE Features for Mobility KPI ................................................................................... 44 Figure 1-45: LTE Features affecting load based mobility(IFLB/IFO/IROW) ................................... 44 Figure 1-46: Automated Mobility Optimization ............................................................................... 46 Figure 1-47: Counters, events and parameters AMO .................................................................... 47 Figure 1-48: Automated mobility optimization................................................................................ 48 Figure 1-49: Counters, events and parameters AMO .................................................................... 48 Figure 1-50: Counters, events and parameters AMO .................................................................... 49 Figure 1-51: Case study 1: AMO Trial test .................................................................................... 49 Figure 1-52: Case Study 1: AMO Feature parameters .................................................................. 50 Figure 1-53: Case Study 1: AMO Performance Results ................................................................ 50 Figure 1-54: Case Study1: AMO Performance Results ................................................................. 51 Figure 1-55: Case Study1: AMO Performance Results ................................................................. 51 Figure 1-56: Case Study 1: AMO Performance Results Conclusion .............................................. 52 Figure 1-57: MCPC: Related Special KPI ...................................................................................... 54 Figure 1-58: Mobility Control at Poor Coverage ............................................................................ 55 Figure 1-59: Mobility Control at Poor Coverage ............................................................................ 56 Figure 1-60: Case 2: Post MCPC evaluation L2600 ...................................................................... 57 Figure 1-61: Case 2: Post MCPC evaluation L2600 ...................................................................... 57 Figure 1-62: Case 2: Post MCPC evaluation L2600 ...................................................................... 58 Figure 1-63: Case 2: Post MCPC evaluation L2600 ...................................................................... 58 Figure 1-64: Case 2: Post MCPC evaluation L1800 ...................................................................... 59 Figure 1-65: Multi-Layer Service-Triggered Mobility ...................................................................... 60 Figure 1-66: Summary of Chapter 1 .............................................................................................. 61 Figure 2-1: Objective of Chapter 2 ................................................................................................ 63 Figure 2-2: Investigation: X2 HO - Preparation.............................................................................. 64 Figure 2-3: Investigation: X2 HO - Execution ................................................................................ 65 Figure 2-4: Investigation: X2 HO - Execution ................................................................................ 66 Figure 2-5: Identifying Events – example ...................................................................................... 67 Figure 2-6: Identifying Events - example ....................................................................................... 67 Figure 2-7: Case 1- IRAT HO, optimum parameter setting to improve LTE drop rate .................... 68 Figure 2-8: Case 1- IRAT HO, optimum parameter setting to improve LTE drop rate .................... 69 Figure 2-9: Case 1- IRATHO, Parameter setting to improve LTE drop rate ................................... 69 Figure 2-10: Case 1- IRAT HO optimum parameter setting to improve LTE drop rate ................... 70 Figure 2-11: Case 2: Oscillating Handover .................................................................................... 70 Figure 2-12: Case 2: Intra frequency HO ...................................................................................... 71 Figure 2-13: Case 2: Oscillating Handover .................................................................................... 71 Figure 2-14: Case 2: Performance result ...................................................................................... 72 Figure 2-15: Case 2: Performance result ...................................................................................... 72 Figure 2-16: Case 3: With AMO HOSR degraded ......................................................................... 73 Figure 2-17: Case 3: CIO statistics of the network ........................................................................ 74 Figure 2-18: Case 3: CIO & distance analysis ............................................................................... 74 Figure 2-19: Case 3: CIO=-2 & -3 distribution In the cluster .......................................................... 75 Figure 2-20: Case 3: Example of unreasonable CIO value............................................................ 75 Figure 2-21: Case 3: Potential reason of unreasonable CIO value ................................................ 76 Figure 2-22: Case 3: Trail suggestion ........................................................................................... 76 Figure 2-23: Case 4: HOSR Execution Improvement On Distance Site......................................... 77 Figure 2-24: Case 4: HOSR Execution improvement with cell range change ................................ 77 - 216 - © Ericsson AB 2017 LZT1381950 R1A Table of Figure Figure 2-25: Case 4: HOSR Execution improvement with cell range changes .............................. 78 Figure 2-26: Case 5: Poor IRAT success rate to poor WCDMA Cell for QCI8 ............................... 78 Figure 2-27: Case 5: Poor IRAT success rate to poor WCDMA Cell for QCI8 ............................... 79 Figure 2-28: Case 5: Poor IRAT success rate to poor WCDMA Cell for QCI8 ............................... 79 Figure 2-29: Poor IRAT HOSR – Other reasons & solutions ......................................................... 80 Figure 2-30: IRAT improvement Solution Procedure Contd… ....................................................... 81 Figure 2-31: Case 6: IRAT improvement Solution Procedure Contd… .......................................... 81 Figure 2-32: Case 6: IRAT improvement Solution Procedure Contd… .......................................... 82 Figure 2-33: Case 6: IRAT improvement Solution Procedure Contd… .......................................... 82 Figure 2-34: Case 6: IRAT improvement Solution Procedure Contd… .......................................... 83 Figure 2-35: Case 7: KPIs Improvement with IFHO parameter setting .......................................... 83 Figure 2-36: Case 7: KPIs (HOSR,Ret.,Thp.) Improvement with ifho parameter setting ................ 84 Figure 2-37: Case 7: KPIs Improvement with IFHO parameter setting .......................................... 84 Figure 2-38: Case 7: KPIs Improvement with IFHO parameter setting .......................................... 85 Figure 2-39: Case 7: KPIs Improvement with IFHO parameter setting .......................................... 85 Figure 2-40: Case 8: Coverage issue, crsGain Tuning .................................................................. 86 Figure 2-41: Case 8: Cell Selection to Increase crsGain ............................................................... 86 Figure 2-42: Case 8: Coverage issue, crsGain Tuning Example ................................................... 87 Figure 2-43: Case 8: Coverage issue, crsGain Tuning .................................................................. 87 Figure 2-44: Case 9: Handover failed during RRC setup............................................................... 88 Figure 2-45: Case 9: Handover failed in preparation phase .......................................................... 88 Figure 2-46: Case 10: PCI Confusion............................................................................................ 89 Figure 2-47: Case 10: PCI Confusion............................................................................................ 89 Figure 2-48: Case 11: ~100% HO Execution failures .................................................................... 90 Figure 2-49: Case 11- RRC HO Execution failures due to PCI Collision ....................................... 90 Figure 2-50: Case 11- RRC HO Execution failures due to PCI Collision ...................................... 91 Figure 2-51: ANR: MOM – ANR Created ...................................................................................... 92 Figure 2-52: Handover event triggered.......................................................................................... 93 Figure 2-53: ANR event triggered ................................................................................................. 94 Figure 2-54: Periodical ANR (Inter-frequency and IRAT HO) ........................................................ 95 Figure 2-55: Parameters EUTRAN 1(3) ........................................................................................ 96 Figure 2-56: Parameters 2(3) ........................................................................................................ 96 Figure 2-57: Parameters 3(3) ........................................................................................................ 97 Figure 2-58: Observability ............................................................................................................. 98 Figure 2-59: ANR Related Counters.............................................................................................. 99 Figure 2-60: Self-Organizing Networks (FAJ 801 0435) ................................................................ 99 Figure 2-61: Automated Neighbour Relation (PCI Conflict Impact) .............................................. 100 Figure 2-62: Automated Neighbour Relation (PCI Conflict Handling) .......................................... 101 Figure 2-63: Automated Neighbour Relation (PCI Conflict Detection DRX) ................................. 102 Figure 2-64: Automated Mobility Optimization (Coordination with ANR) ...................................... 103 Figure 2-65: CSFB ...................................................................................................................... 104 Figure 2-66: CSFB – Radio feature list........................................................................................ 104 Figure 2-67: KPI Definition Example ........................................................................................... 105 Figure 2-68: CSFB Cases for long CALL SETUP TIME .............................................................. 105 Figure 2-69: CSFB Case 1: Longer setup time due to Core Network signaling ........................... 106 Figure 2-70: CSFB Case 1: CSFB failure due to long Cell reselection time ................................. 106 Figure 2-71: CSFB Case 1: SIB reading during Cell reselection.................................................. 107 Figure 2-72: CSFB Case 1: Failure due to Multiple RRC connection Request ............................ 107 Figure 2-73: CSFB Case 1: Failure due to Multiple RRC connection Request ............................ 108 Figure 2-74: CSFB Case 1: Failure due to Bad UTRAN Coverage.............................................. 108 Figure 2-75: Case 2: UE Fail to Return to LTE After CSFB to UTRAN ........................................ 109 LZT1381950 R1A © Ericsson AB 2017 - 217 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop Figure 2-76: Case 2: UE Fail to Return to LTE After CSFB to UTRAN ........................................ 109 Figure 2-77: Case 2: UE Fail to Return to LTE After CSFB to UTRAN Answer ........................... 110 Figure 2-78: Case 3: UE took more time when return back to LTE NW1 ..................................... 111 Figure 2-79: Case 3: UE took more time when return back to LTE NW2 ..................................... 111 Figure 2-80: Case 3: UE took more time when return back to LTE NW3 ..................................... 112 Figure 2-81: Summary of Chapter 2 ............................................................................................ 113 Figure 3-1: Objective of Chapter 3 .............................................................................................. 115 Figure 3-2: Review: EUTRAN Throughput KPIs .......................................................................... 116 Figure 3-3: Review: EUTRAN Throughput KPIs .......................................................................... 117 Figure 3-4: Review: DL DRB Traffic Volume Measurement ......................................................... 117 Figure 3-5: Review: DL DRB Traffic Volume Counters ................................................................ 118 Figure 3-6: Review: DL DRB Traffic Volume Counters ................................................................ 119 Figure 3-7: Review: EUTRAN Throughput KPIs .......................................................................... 119 Figure 3-8: Review: UL DRB Traffic Volume Measurement ......................................................... 120 Figure 3-9: Review: UL DRB Traffic Volume Counters ................................................................ 120 Figure 3-10: Review: EUTRAN Latency KPIs.............................................................................. 121 Figure 3-11: Review: Downlink Latency Measurement ................................................................ 122 Figure 3-12: Downlink Latency Counters .................................................................................... 122 Figure 3-13: EUTRAN Packet Loss KPIs .................................................................................... 123 Figure 3-14: EUTRAN Packet Loss KPIs .................................................................................... 124 Figure 3-15: Downlink Packet Loss Measurement ...................................................................... 125 Figure 3-16: Uplink Packet Loss Measurement ........................................................................... 125 Figure 3-17: Packet Loss Counters ............................................................................................. 126 Figure 3-18: Packet Loss Counters ............................................................................................. 126 Figure 3-19: Packet Loss Counters ............................................................................................. 127 Figure 3-20: Throughput Optimization ......................................................................................... 127 Figure 3-21: Analysis Flow for Dl Throughput Investigation ......................................................... 128 Figure 3-22: Analysis Flow for UL throughput Investigation ......................................................... 129 Figure 3-23: Downlink Throughput Optimization ......................................................................... 129 Figure 3-24: Downlink Throughput Optimization (contd..)............................................................ 130 Figure 3-25: Uplink Throughput Optimization .............................................................................. 130 Figure 3-26: Uplink Throughput Optimization (contd..) ................................................................ 131 Figure 3-27: Throughput Testing: Checks ................................................................................... 131 Figure 3-28: PRB Utilization ....................................................................................................... 132 Figure 3-29: PRB Utilization ........................................................................................................ 133 Figure 3-30: Ways to reduce PRB utilization ............................................................................... 133 Figure 3-31: Essential Parameters .............................................................................................. 134 Figure 3-32: Counters for throughput analysis ............................................................................ 134 Figure 3-33: RSSI ....................................................................................................................... 135 Figure 3-34: UL SINR ................................................................................................................. 135 Figure 3-35: UL RLC NACK ........................................................................................................ 136 Figure 3-36: Power restricted transport block-UL ........................................................................ 137 Figure 3-37: CQI ......................................................................................................................... 137 Figure 3-38: MIMO RANK distribution usage .............................................................................. 138 Figure 3-39: Modulation scheme usage DL ................................................................................. 138 Figure 3-40: Modulation scheme usage UL ................................................................................. 139 Figure 3-41: Throughput Optimization-essential checks .............................................................. 139 Figure 3-42: QoS Framework ...................................................................................................... 141 Figure 3-43: QoS Aware Scheduler – Absolute priority example ................................................. 142 Figure 3-44: Scheduler Configuration.......................................................................................... 143 Figure 3-45: Proportional Fair Scheduling Benefits ..................................................................... 144 Figure 3-46: LTE Features enhancing UL/DL Throughput ........................................................... 144 - 218 - © Ericsson AB 2017 LZT1381950 R1A Table of Figure Figure 3-47: 256 QAM DL Parameters ........................................................................................ 146 Figure 3-48: Link Adaptation Tables............................................................................................ 148 Figure 3-49: 64-QAM Uplink (FDD/TDD) ..................................................................................... 149 Figure 3-50: 64 QAM UL Benefits ............................................................................................... 149 Figure 3-51: 64 QAM UL Parameters .......................................................................................... 151 Figure 3-52: Ericsson Lean Carrier Applying 5G concepts to today’s 4G LTE ........................... 152 Figure 3-53: Ericsson Lean Carrier Applying 5G concepts to today’s 4G LTE ........................... 154 Figure 3-54: 4CC DL Carrier Aggregation Extension ................................................................... 155 Figure 3-55: Uplink Carrier Aggregation ...................................................................................... 156 Figure 3-56: Uplink Carrier Aggregation Parameters ................................................................... 157 Figure 3-57: Advanced Carrier Aggregation (FAJ 801 0568) ....................................................... 158 Figure 3-58: Inter-eNB Carrier Aggregation................................................................................. 159 Figure 3-59: Inter-eNB Carrier Aggregation Performance ........................................................... 160 Figure 3-60: Inter-eNB Carrier Aggregation Parameters ............................................................. 161 Figure 3-61: Multi-carrier Load Management (FAJ 801 0424) ..................................................... 162 Figure 3-62: Prescheduling ......................................................................................................... 162 Figure 3-63: Prescheduling OFF ................................................................................................. 163 Figure 3-64: Prescheduling ON ................................................................................................... 164 Figure 3-65: Prescheduling Parameters ...................................................................................... 164 Figure 3-66: LTE transmission modes Downlink transmission modes ......................................... 167 Figure 3-67: Which MIMO Mode is Best? .................................................................................... 167 Figure 3-68: VOLTE Impact on BB Throughput & DBS ............................................................... 168 Figure 3-69: VOLTE Impact on BB Throughput & DBS ............................................................... 169 Figure 3-70: VOLTE Impact on BB Throughput & DBS ............................................................... 169 Figure 3-71: Summary of Chapter 3 ............................................................................................ 170 Figure 4-1: Objective of Chapter 4 .............................................................................................. 171 Figure 4-2: CASE 1: DL Throughput ........................................................................................... 172 Figure 4-3: CASE 2: UL Throughput Improvement ...................................................................... 173 Figure 4-4: Case 3: LTE Latency issue ....................................................................................... 174 Figure 4-5: Case 3: Investigation Analysis .................................................................................. 175 Figure 4-6: Case 3: Investigation Analysis .................................................................................. 175 Figure 4-7: Case 3: Investigation Analysis .................................................................................. 176 Figure 4-8: Case 3: Investigation Analysis .................................................................................. 177 Figure 4-9: Case-Study 4: Reported issue .................................................................................. 179 Figure 4-10: Exercise 1 – Analyze ROP files............................................................................... 180 Figure 4-11: Exercise 1 - Answer ................................................................................................ 180 Figure 4-12: Exercise 2: .............................................................................................................. 181 Figure 4-13: Exercise 2 - Answer ................................................................................................ 181 Figure 4-14: Exercise 2 – Answer ............................................................................................... 182 Figure 4-15: Exercise 2 – Answer cont’d ..................................................................................... 182 Figure 4-16: Exercise 2 – Answer cont’d ..................................................................................... 183 Figure 4-17: Exercise 2 – Answer Cont’d .................................................................................... 183 Figure 4-18: Exercise 2 – Answer Cont’d .................................................................................... 184 Figure 4-19: Exercise 2 – Answer Cont’d .................................................................................... 184 Figure 4-20: Exercise 2 – Answer Cont’d .................................................................................... 185 Figure 4-21: Exercis 2 – Answer Cont’d ...................................................................................... 186 Figure 4-22: Exercise 2 – Answer Cont’d .................................................................................... 186 Figure 4-23: Exercise 3: .............................................................................................................. 187 Figure 4-24: Exercise 3 – Answer Cont’d .................................................................................... 187 Figure 4-25: Exercise 3 – Answer Cont’d .................................................................................... 188 Figure 4-26: Exercise 3 – Answer Cont’d .................................................................................... 188 LZT1381950 R1A © Ericsson AB 2017 - 219 - LTE Mobility and Throughput - KPI Analysis & Optimization Workshop Figure 4-27: Exercise 3 – Answer Cont’d .................................................................................... 189 Figure 4-28: Exercise 3 – Answer Cont’d .................................................................................... 190 Figure 4-29: Exercise 3 – Answer Cont’d .................................................................................... 190 Figure 4-30: Exercise 3 – Answer Cont’d .................................................................................... 191 Figure 4-31: Exercise 3 – Answer Cont’d .................................................................................... 191 Figure 4-32: Exercise 3 – Answer Cont’d .................................................................................... 192 Figure 4-33: Exercise Summary .................................................................................................. 192 Figure 4-34: WAY FORWARD .................................................................................................... 193 Figure 4-35: Summary of Chapter 4 ............................................................................................ 194 - 220 - © Ericsson AB 2017 LZT1381950 R1A