Huawei Technologies, Canada Ltd.
Contact: gamini.senarath@huawei.com
WINLAB Fall 2015 Conference
Dec. 03-04, 2015
NJ, USA
Huawei proprietary

500+ 5G Experts
Stockholm
Ottawa
New Jersey
Paris
 Munich Moscow
Shanghai
Chengdu
Shenzhen
9 5G Research Centers
Stockholm, Sweden
• System Architecture
• Algorithms
 Munich, Germany
• Verticals
Paris, France
• Standardization
Moscow, Russia
• Fundamental
Algorithms
New Jersey, USA
• 5G Transmission
 Ottawa, Canada
• 5G Radio
• Network Architecture
5G Research Centers in China
• Shen zhen
• Shang hai
• Cheng du





Huawei proprietary Page 3

Huawei proprietary Page 4

5G will enable new applications , new business models , and even new industries
Traffic
/Revenue
Video
Web
10Gbps
Throughput
1ms
Latency
1 Million
Voice Vehicular
Telematics
Massive IoT
AR/VR
MirrorSys High Speed
Railway
Teleoperation
Connections/km 2
……
Use Cases
Body Today's Long Tail
Huawei proprietary Page 5

Latency
ms
E2E
Latency
Throughput Connections
bps
Per
Connection
Connections
Per km 2
Mobility
km/h
High-speed
Railway
Network
Architecture
Ability
Required
30~50x
30~50ms
100x
100Mbps
100x
10K
Huawei proprietary
1.5x
350Km/h
NFV/SDN
Inflexible
Page 6

Enhance Mobile
Internet
Empower
Internet of Things
7
Customers (Verticals es &
Other Network Providers )
• Easy access to the common infrastructure of 5G
• Real-time, on-demand service
Network
Providers
• Easy deployment & maintenance
• Flexibility for multiple industries (SLICING)
Infrastructure
Providers
• Combine infrastructure to form one infrastructure for network providers
End User groups
• Ubiquitous consistent experience
• New services

2014 2015 2016
 …  R12  R13  R14
 5G innovations will be applied to
4G
5G
2017
 R15
 4
G
 4G will simulate the emergence of new applications for
5G

2018
 R16
2019
 …
 Revolution
 Evolution
Huawei proprietary Page 8

Spectrum
Aggregate All
Available Bands
New Network
Architecture
One Physical Network
Multiple Industries
Huawei proprietary
New Air Interface
Flexibility &
Spectrum Efficiency
Page 9

WRC15
Requirement >500MHz for IMT-2020
WRC19
45GHz available for future Cellular Access and Self-Backhaul
Cellular
Bands
Visible
Light
1 2 3 4 5 6 10
5G Primary bands
20 30 40 50 60 70 80 90 100
GHz
5G Complementary Bands for Capacity, 45GHz available
Huawei proprietary Page 10

Huawei proprietary Page 11

Diverse
Applications
Voice Web Video Verticals……
UAI
(Unified Air Interface) to meet the diverse requirements
Diverse
QoE
Diverse
Adoption
Data Rate Latency Connections Battery Life Outdoor/ indoor
Wide/Deep coverage
Low/High band
Wide/Narrow
Bandwidth
Page 12 Huawei proprietary

Full Duplex Massive MIMO
Mobile Internet
Adaptive
Air Interface
Internet of Things
SCMA Polar Code
F-OFDM
SCMA: Sparse Code Multiple Access
F-OFDM: Filtered OFDM
One air interface fits many
applications with high flexibility, at least a
3x spectrum efficiency improvement
Page 13

Cell centric cellular to UE centric no cell C-RAN based UE centric TP Optimization
D2D enabled UE Cooperation

Abstraction of the UE radio access with virtualized the cell concept to enable RAN slicing by
 Decoupling the UE from physical cell-site

Decoupling DL/UL
 Decoupling Control/Data

Decouple physical topology with services
 New UE and network transmit node association mechanism enabled by “Hyper cell ID” and “Dedicated
UE connection ID”
 CRAN and D2D enabled UE centric transmission point
(TP) cooperation and device cooperation to eliminate
“cell edge”
 New UE states support massive connected devices with low signaling overhead and energy consumption
 Seamless mobility transparent to UE with simplified procedural and reduced latency
Huawei proprietary













New waveform e.g. f-OFDM
Wider Bandwidth
Adaptive frame structure
Non-Orthogonal Multiple Access, e.g. SCMA
UCNC
Massive MIMO
Polar Code eMBB
Enhanced Mobile
Broadband
Grant-free multiple access
Narrow band SCMA
Asynchronous (TA-free) Transmit
UE dedicated connection ID
Polar Code for small packet mMTC
Massive Machine Type
Communications
Future IMT







Shorter TTI
SCMA based grant-free Tx
Fast system re-entry scheme
ACK/NACK less re-transmission
UE cooperation diversity
New data notification methods
Polar Code uMTC
Ultra-reliable and Low-latency
Communications
Huawei proprietary

Page 16 Huawei proprietary

5G 3G/4G
1.
Infinite types of services/applications with huge disparate QoE/QoS requirements are emerging
1.
One-fit-all user plane architecture not optimal
•
Different service requires different mobility management, charging policy, authentication, etc
2.
MME, RRC, PCRF, etc only optimized for individual mobile services
2.
Operation of network should be optimized for different vertical services
3.
Openness of future networks – service customized functions, other
3.
Closed model in 3G/4G than network functions
4.
Integration of eMBB/mMTC/uMTC
Huawei proprietary
4. By nature, 3G/4G optimized for personal communications
Page 17

(Service/Infrastructure)
NFV and SDN
Huawei proprietary Page 18

End User
Population
(Customer 1)
End User
Population
(Customer 2)
VN Customer 1
Slice
VN Customer 2
Slice
Infra-structure Abstraction
Also called VNO (VN operator)
Network
Provider A
Network
Provider B
Control + Resources
Or
Connectivity service
(Dynamic or static)
Infra-structure
Provider C
Infra-structure
Provider A
Access Points
VN Customer has an end user
(device) population

MTC (Alarm, Sensor company)

Video distribution company
 Police, Fire
 E-health monitoring service
Network Provider  Telecom
Connectivity Service Provider
(TCSP)
Infra-structure Provider (InP)
 Own or borrow resources from InPs.
 Service Provider A may own infrastructure Y

Provide resources and controlling technology with Phy abstractions
 Dynamic or static
 Provide connectivity service in specific geographical area
TCSP Offers an E2E Service Slice to the Customer
Huawei proprietary

Edge network segment
• hard slicing
Central network segment
• soft slicing
3G/4G network
• a network slice
8888
8888
8888 eMBB slice
D2D slice
eMTC slice
cMTC slice
Other slices (common or
4G slice)
5G Key L1
Enabling
Technologies
20
Physical
DC
D-RAN C-RAN
DC DC
@
GWNI

(Independent, Isolated, E2E)
Slice-1
Network Slicing Technologies
(Examples)
Slice-2
1. Dynamic integer programming
Slice-4
Slice-3 algorithm for fast network topology generation
2. Minimum perturbation reoptimization linear programming algorithm
Drastically Reduce the Dependence of
Network Functions
21

 Service Request by VN operator (with Service attributes)
 E.g., service function chain, Transport and Traffic distribution (time and space)
 Admission Control ( VNAC )
 Creation of a Slice instance ( Software Defined Topology – SDT )
 Only Virtual Topology with instantiated VNFs
 Network with reserved resources may allocate physical resources
 Slice Operation ( Software Defined Radio Resource Allocation –
SDRA)
 Traffic engineering , monitoring, policing, charging etc.
 Slice Termination
Huawei proprietary

Network Provider 2
Network Provider 1
Network Provider
Information
Database
(Public)
OSS/BSS
Global Customer
Service Mgmt (CSM)
Orchestrator/SONAC
SDRA
- VNAC
SDT SDP
VNFM VIM
Customer
(e.g. A vertical service operator)
CSM
Slice
A
Connectivity service provided by Slice A
Connectivity service provided by Slice B
Common Control functions
(network controlled)
Slice Specific control functions
(network or customer controlled)
Slice specific user plane functions (network or customer controlled)
Infrastructure
VNFs
Control Functions (customer controlled) User Plane Functions (customer controlled)
Control Functions (network controlled)
Huawei proprietary
User Plane Functions (network controlled) e.g. Per user
Virtual SGW
Per service Virtual
SGW

A service slice is specifically prepared for the customer by slicing the network, i.e., by creating a Network Slice. Several options:
 Use existing matching network slice descriptor
- E.g. Another operator requests a eMBB slice. It is created using existing eMBB description.
 Introduce a new network slice descriptor to create a network slice instance
- E.g. Vertical service operator requests a new service (e.g. CDN). It is created using customized network slice descriptor.
 Integrate into an existing network slice instance, e.g. using same resources, e.g.:
Another MTC operator requests a similar MTC service
An operator already having a network slice requests another service using the same resource pool
Huawei proprietary

(Service Oriented Network-Slice Auto-Creator) - Example
U1
Service level Graph
QoE/QoS requirement
Customer Service Description/requirement
U2
VNAC
+SDT
SDRA
Logical topology mapping to physical network resource
SDP
End-to-end transport protocol design
VN graph
U1
V-u-SGW (U1) F1 V-u-SGW (U2)
Forwarding graph description; link description
VN logical topology (placement of v-s-SGWs in infrastructure)
U1
WN infrastructure (resource pool)
VN physical topology
U1
VN with customized protocol
U1
U2
U2
U2
Transport protocol defined by SDP
U2
25

VN Admission/service negotiation:
–
Different customer would have different demand distributions in time and space
– How to get multiplexing gain
– Different services needed different amount of resources based on QoS and geographical distribution
– How the charging is done for a customer having a large number of users with different services
–
When customer request multiple slices using same resources how to make the admission decision
Software Defined Topology (Virtual if resources are not reserved)
– Optimal placement of the Service Functions? Virtual Topology Placement depend son traffic, mobility.
• Fast moving user can have its SGW much inside the network while slow moving user can have its SGW close to edge. Similarly caching functions.
– For a hard slice or resource reserved slice physical topology also established.
–
E.g., MTC type of services aggregation points, message filtering, Customer functions , should be
26 Page 26 strategically placed.

•
Traffic Engineering (TE)
–
Slice Specific TE does dynamic resource allocation to slices and sessions. Per slice KPI and QoE guarantees are needed.
27
– If resource sharing is done, Global traffic engineering is required across the slices
–
Depending on traffic load, invokes a resource coordination function for local areas, take action to control traffic or trigger for service re-negotiation
– QoE Guarantee – using user’s feedback or action monitoring, QoS to QoE mapping tables are stored (per user based, per group based, per application based) to deliver required QoS
–
Access schemes for massive MTC and efficient short packet delivery
•
Customer Service management, Connectivity Management, Caching and
Pre-fetching, Context Management
– Imbalance between Demand vs Revenue prediction curves (e.g., Demand based charging , User in the loop )
–
Per user and per service/slice based mobility handling/tracking
– Per slice based content distribution and caching and pre-fetching based on per user/group
–
Context data analytics/storage and using them for efficient service delivery and for 3 rd party usage

Huawei proprietary

Industry : NGMN, 5GPPP, METIS II, FANTASTIC 5G, mmMAGIC,
5GXhaul, 5G-EX
Standards: ETSI /NFV , 3GPP (SA1, SA2 and SA5)
Current/recent activities in NGMN
– White paper on 5G issued in January, 2015
• (a) 5G vision; (b) 5G requirements; (c) 5G Architecture concepts; (d) Spectrum considerations;
• Basis for many other 5G standard organization activities
Currently four work streams under Project P1
• WS1-Architecture, WS2-Verticals, (3) WS3-Requirements for better MBB and Telco services and, (4) WS4- Interacts with standard development organizations.
• WS1Further work under 3 groups:
• E2e Architecture – currently discuss definition of the SLICE
• Network and Service Management
• Security
Huawei proprietary Page 29

Current/recent activities in 3GPP
SA1 – Services
• Discussed 5G use cases and categorized them into 4 main areas
− Enhanced Mobile Broadband (eMBB): higher capacity; enhanced connectivity; higher user mobility.
− Critical Communications (CriC): higher reliability with lower latency; higher reliability and higher availability with lower latency; very low latency; higher accuracy positioning.
− Massive Internet of Things (mIoT): high connection density; low complexity; low power consumption.
− Network Operations (NEO): flexible functions and new value creation; migration and interworking; optimizations and enhancements.
SA2 – Architecture: Currently discusses the following
- Key NextGen Architecture Requirements
- Key Technical Areas and Key issues that need to be addressed
Huawei proprietary Page 30

2010 2011 2012 2013
WRC-12
2014 2015 2016
WRC-15
2017 2018 2019 2020 2021
WRC-19
2022
ITU-R
ITU Workshop
Req., Eval.
Criteria
Proposal
Eval
Spec.
RAN
3GPP
Rel-10 Rel-11 Rel-12 Rel-13
We are here
5G WI(s)
 Phase-1: fundamental features of UAI focusing on spectrum below 6GHz
 Phase-2: enhancement features of UAI below and above 6GHz
Rel-14 Rel-15 Rel-16
5G SI(s)
UAI, other features / enhancements
Rel-17
5G SI(s)
 Start from UAI below 6GHz
 UAI above 6GHz will follow up after the channel model above
6GHz is ready
5G WI(s)
Phase-1
5G WI(s)
Phase-2
LTE-Advanced (4G) LTE New Branding (4.5G)
Notes:
* Proposal submission to ITU no later than June 2019
* Spec submission to ITU no later than February 2020 Huawei proprietary

Copyright©2015 Huawei Technologies Co., Ltd. All Rights Reserved.
The information in this document may contain predictive statements including, without limitation, statements regarding the future financial and operating results, future product portfolio, new technology, etc. There are a number of factors that could cause actual results and developments to differ materially from those expressed or implied in the predictive statements. Therefore, such information is provided for reference purpose only and constitutes neither an offer nor an acceptance. Huawei may change the information at any time without notice.
Huawei proprietary Page 32

Number of Connections
375 400
350
300
250
200
150
100
50
0
95 x3.9
LTE-Advanced SCMA

Non-orthogonal multiplexing of layers

Overloading to increase overall rate and connectivity

Sparsity to limit complexity of detection

Multi-dimensional codewords with shaping gain and better spectral efficiency

Spreading for robust link adaptation

Grant-free access for reduction of both latency and signaling overhead
Huawei proprietary

 For small packet (e.g. IoT, control channel), Polar Codes have 0.5-2dB gain comparing with
Turbo Code used in LTE, the gain is significant.
 No error floor, suitable for ultra-reliable transmission
 Low energy consumption
Huawei proprietary

 Optimized RAT for each application/use case
 Dynamic or semi-static or static configurable
 Across frequency carriers or within the same frequency carrier
 Forward compatible: easy to add future-proof new service/use case
 Smooth migration of LTE
One size fits all (LTE)  Air Interface Adaptation
(5G)
Soft AI
Waveforms and MA
 f-OFDM

SCMA
 Ultra NB WF
Coding Modulation

Polar

Turbo
 Network Coding
Access Protocols

Scheduled

Grant-free
 Adaptive HARQ
Frame Structure
 Flexible TTI

Flexible
Numerologies
 Flexible Duplex
 Full-Duplex
Huawei proprietary

Same Carrier
Filtered-OFDM
Flexible subcarrier parameterization uMTC MBB mMTC
Numerology-1 Numerology-2 Numerology-3
 Enable future proof design and RAN slicing by allowing independent co-existence of multiple services within the same carrier
 Sub-band digital filter to control inter-block interference (spectrum localization)
 Orthogonal Intra block to maintain OFDM benefits
 Non-orthogonal to enable co-existence of multiple numerologies without guard band
Huawei proprietary
MBB Smart Metering Driverless Car Broadcast / Multicast
Frequency