IEEE 802.16ppc-10/0049r1
Project
IEEE 802.16 Broadband Wireless Access Working Group <http://ieee802.org/16>
Title
Functional Requirements for Network Entry and Random Access by Large Number of
Devices
Date
Submitted
2010-08-25
Source(s)
N. Himayat, K. Johnsson, S. Talwar, Xinrong Wang
E-mail: nageen.himayat@intel.com
Intel Corp.
Re:
Requirements for network entry and random access for large number of devices
Abstract
This contribution analyzes functional requirements for network entry and random access
procedures, which result from a large number of devices connecting to the network.
Purpose
For review and adoption into the IEEE 802.16p Systems Requirements Document
Notice
Release
Patent
Policy
This document does not represent the agreed views of the IEEE 802.16 Working Group or any of its subgroups. It
represents only the views of the participants listed in the “Source(s)” field above. It is offered as a basis for
discussion. It is not binding on the contributor(s), who reserve(s) the right to add, amend or withdraw material
contained herein.
The contributor grants a free, irrevocable license to the IEEE to incorporate material contained in this contribution,
and any modifications thereof, in the creation of an IEEE Standards publication; to copyright in the IEEE’s name
any IEEE Standards publication even though it may include portions of this contribution; and at the IEEE’s sole
discretion to permit others to reproduce in whole or in part the resulting IEEE Standards publication. The
contributor also acknowledges and accepts that this contribution may be made public by IEEE 802.16.
The contributor is familiar with the IEEE-SA Patent Policy and Procedures:
<http://standards.ieee.org/guides/bylaws/sect6-7.html#6> and
<http://standards.ieee.org/guides/opman/sect6.html#6.3>.
Further information is located at <http://standards.ieee.org/board/pat/pat-material.html> and
<http://standards.ieee.org/board/pat>.
1 Introduction
This contribution analyzes functional requirements for network entry and random access procedures, which
result from a large number of devices connecting to the network. The smart metering application is analyzed in
greater detail to highlight key issues. We focus on initial network entry using ranging channels for random
access. However general conclusions are also applicable to other random access channels as well.
Traffic characteristics and high level requirements for Smart Metering applications were discussed in [1]. Initial
network entry procedures are applicable for several smart metering scenarios:
1. Initial network entry
2. Network re-entry after power-outage
3. Use of network entry procedure for short message transmission instead of establishing full connection
given that smart metering is for most cases a time-tolerant application with typically intermittent traffic.
In this contribution we characterize the excess load caused by a large number of smart meters performing initial
1
IEEE 802.16ppc-10/0049r1
network entry and investigate the capacity of the current IEEE 802.16m standard [2] to address the excess load.
The resulting requirements are captured as modifications to the IEEE 802.16p System Requirement Document
(SRD) [4].
2 Estimating Number of Devices (Smart Metering)
In this section we estimate typical number of devices per sector requiring network connectivity by using the
smart metering application as an example. Some assumptions used in the calculations are as follows:
• Uniform population density
• 3 meters (gas, electric, water) per house
• 3 sectors/cell
• 100% penetration
• Washington D.C statistics 334K users / 73 sq. km
• NYC statistics: 10K users /sq. km
• *Urban London example based on Vodafone’s estimates [3].
Scenario
Population
Density
Household
Size
Cell Sizes
Max.
Meters/Sector
Urban
(New York City)
1 SS/100 sq.
meters
2.6
500m/1km
3000/12,000
Suburban
(Washington
D.C example)
1 SS/220 sq.
meters
2.6
1 Km/2 km
5500/11,000
Urban London*
1 SS/133 sq.
meters
2.64
1Km/2Km
8943/35,771
Table 1: Estimated Number of Meters/Sector for representative geographies as a function of cell size.
It can be seen that there can be a large variation in the number of M2M devices per sector depending on the
demographics and type of cellular deployments.
3 Support for Ranging Channels in IEEE 802.16m
In this section we analyze the capacity of random access channels in IEEE 802.16m. Various ranging channel
configurations are described in [2]. Here we assume a representative case using the following set of
assumptions:
•
•
•
•
•
Frame structure corresponding to 5 MHz bandwidth with CP=1/4
Normal ranging format with each ranging channel allocation equaling 1 sub-band x 1 sub-frame (6
symbols)
Cell sizes supporting the maximum allowed number of random access preambles per channel (32 codes
for initial network entry, minimum of 8)
Allowed arrival rate to support 1% contention probability per preamble (assuming Poisson arrival, see
Appendix)
Cases covering different periodicities of ranging channel (ranging from once every frame to once every 4
super-frames)
2
IEEE 802.16ppc-10/0049r1
Case Number of
Contention
Channels
/second
1
200
2
50
3
25
4
12.5
Maximum
(Minimum)
# of Codes
Random Access
Opportunities
/second
Overhead
32 (8)
32 (8)
32 (8)
32 (8)
6400
1600
800
400
2.85%
0.71%
0.35%
0.18%
(1600)
(400)
(200)
(100)
Number of Users/s
w/ contention probability
of 1%
(0.15 arrival rate)
960 (240)
240 (60)
120 (30)
60
(15)
Table 2: Maximum number of users per second, who are able to perform random access with a one
percent contention probability per preamble.
4 Ranging Channel Access Rates for Smart Metering Applications
Table 3 analyzes the typical random access rates per second for various smart metering traffic scenarios,
assuming dense deployments representative of the U.S and U.K environments. We consider traffic scenarios
related to a) meter readings at various reporting intervals b) alarm reporting and message acknowledgements c)
network access after power restoration, given unsynchronized reporting by the meters.
For this calculation it is assumed that the random access attempts are uniformly distributed across the interval of
interest. It can be seen that for cases where random access attempts are not regulated and are unsynchronized,
the network access attempts may exceed the maximum capacity supported by IEEE 802.16m, for reasonable
overhead allocation for ranging channels.
Application
Interval of Interest
Random Access
Attempts/second
(12K devices/sector)
Random Access
Attempts/second
(35K devices/sector)
Reporting
5 minutes
40
116.7
Reporting
15 minutes
13.33
38.9
Alarms
1 minute
200
583.3
Unsynchronized
meters
10 seconds
600
3500
Table 3: Ranging channel access rates for various smart metering traffic scenarios.
5 Requirements for Network Entry and Random Access with Large Number of
3
IEEE 802.16ppc-10/0049r1
Devices
Based on the data included in earlier sections, we propose that IEEE 802.16p system should address the
following requirements.
1. Improved Regulation of Network Access Attempts Network entry attempts by devices should be
regulated to avoid simultaneous access by a large number of devices. Here we note that IEEE
802.16 standards support self-regulated access based on random back-off procedures given
contention on random access channels. However, such procedures will result in added latency,
especially if the number of devices contending is large. Additionally, it is desirable to minimize the
random access overhead in the system and to reduce the number of channels dedicated for random
access. This tradeoff between the random access latency introduced as a function of number of
devices attempting network entry for a fixed number of random access channels is FFS.
2. Support for Prioritized Network Access Attempts The system should be able to prioritize network
entry by certain devices or traffic classes that are not delay tolerant, so that they are not impacted by
the presence of large number of devices attempting network access simultaneously.
3. Increased random access channels for network access The system should also improve the
number of preambles available for random access by increasing the number of sub-bands available
for contention. It should also provide for flexible configuration of random access partitions. For
example for M2M use cases based on stationary devices, the number of codes assigned for handoff
may be re-assigned for initial network entry.
4. Reduction in Perceived Network Load through the use of Aggregation Devices The system may
also consider the use of aggregation devices to reduce network load.
5. Improved Short Message Service Given that delay tolerant applications may minimize the time
spent in connected stated and therefore utilize the SMS (short message service) service via initial
ranging channel to transmit short bursts of data, it may be desirable to provide flexibility in the SMS
message size to minimize the number of random access attempts for SMS transport.
Performance requirements related to the allowed network latency as a function of the intensity of random access
for a given number of random access channels configured are FFS.
6 Text Proposal
Modify the functional requirements described in Section 6.2 of [4] as shown.
---------------------------Begin Text Proposal------------------------------6.2
Large Numbers of Devices
The 802.16p amendment should support very large numbers of devices.
4
IEEE 802.16ppc-10/0049r1
6.2.1
The protocol should be able to address large number of devices individually or by group.
6.2.2
The protocol should support group management for large numbers of M2M devices dispersed over a
large area.
6.2.3
The 802.16p system should support network access from large numbers of M2M devices.
 The system should provide improved regulation of network access to avoid contention delays caused
from simultaneous network access by large number of devices.
 The system should allow for greater flexibility in assigning sub-bands for initial network entry.
 The system should support greater flexibility in assigning ranging preambles for initial network
entry.
 The system may make use of aggregation devices to reduce the number of devices requesting random
access.
6.2.4
The 802.16p system should provide priority access for M2M devices across different classes of M2M
services.
 The system should prioritize network entry based on M2M device or flow type.
6.2.5
The 802.16p amendment should support large numbers of M2M devices in connected state as well as in
idle state.
6.2.6
The 802.16p system should support QoS for M2M services.
6.3 Small Burst Transmissions
The standard should support very small burst transmissions.
6.3.1 The 802.16p amendment should support efficient transmission of small burst sizes.
 The system should improve the efficiency of the short message service for efficient transmission of
small bursts.
 The system should allow greater flexibility in configuring the message size defined for ranging
channel based short message service, in order to minimize the number of random access attempts
required to send the message.
-----------------------------------------End Text Proposal-----------------------
7 References
1. N. Himayat et. al, “Smart grid requirements for IEEE 802.16 M2M networks,” IEEE C802.16ppc10_0042r2, July 2010.
2. IEEE P802.16m/D7
3. Vodafone, “RACH intensity of time-controlled devices,” R2-102296, 3GPP TSG RAN WG2 #69bis.
4. IEEE 80216ppc-10_0011, “IEEE 802.16p Machine to Machine (M2M) System Requirements
Document (SRD),” initial working document, July, 2010.
5
IEEE 802.16ppc-10/0049r1
8 Appendix
The figure bellows plots the contention probability for a given random access preamble as a function of the
arrival intensity of users. A Poisson arrival model with parameter lambda is assumed. The contention
probability is given by the probability that more than 1 users contend for a given preamble and is given by
Pc = 1 – exp(-lambda) – lambda exp(-lambda)
The average arrival rate per preamble may be obtained by dividing the number of random access attempts by the
total number of random access preambles. The figure below indicates that for a 1% collision probability the
arrival rate must be maintained below 0.15.
0
10
-1
prob of collision
10
-2
10
-3
10
0
0.2
0.4
0.6
0.8
1
lambda
1.2
1.4
1.6
1.8
2
Figure 1: Probability of collision as a function of the arrival rate (lambda) of random access attempts
assuming Poisson arrival rate.
6