Sunteți pe pagina 1din 21

5G RAN: Design Guidelines

and Key Considerations


Presenter: Ömer Bulakci (Huawei Tech. GRC), WP5 Leader

5G PPP Workshop on 5G Architecture and RAN Integration

2017-03-30
METIS-II Key 5G Architecture Paradigms
Moving
5G RAN – a
A Logical CN/RAN Functionality from
Harmonized and
Split with evolved CN to RAN:
Integrated
Interfaces Common RAN CP
Landscape of AIVs
Framework

Functionality on a
RAN Protocol
Faster Time Scale:
Stack RAN Enablers for
Agile Traffic
Considerations Network Slicing
Steering and
(Functional Splits)
Resource Mgmt
Covered in this presentation Not covered herein
Moving Functionality from Core Network to RAN
Common RAN Control Plane (CP) framework
Services & Applications

› New more efficient common S1*


S1*
4) Mobility,
initial access (RACH and 5) D2D

Assign services to resources


LTE-NR tight communication
paging) integration 1) UE in IDLE or
INACTIVE mode,
camping
› Robust mobility to support LTE
all 5G scenarios, 1-100 Zzz
GHz NR z
X2*
X2*

› Tight integration with LTE NR


2) Initial access, UE
3) Mobility, enters CONNECTED
NR
› D2D communication multi-
connectivity, BF
state

integral part of the RAN mobility


design

› Agile RM Framework
METIS-II, Architecture WG, Page 3
Moving Functionality from Core Network to RAN
Common RAN Control Plane (CP) framework
Services & Applications

› New more efficient common S1*


S1*
4) Mobility,
initial access (RACH and 5) D2D

Assign services to resources


LTE-NR tight communication
paging) integration 1) UE in IDLE or
INACTIVE mode,
camping
› Robust mobility to support LTE
all 5G scenarios, 1-100 Zzz
GHz NR z
X2*
X2*

› Tight integration with LTE NR


2) Initial access, UE
3) Mobility, enters CONNECTED
NR
› D2D communication multi-
connectivity, BF
state

integral part of the RAN mobility


design

› Agile RM Framework
METIS-II, Architecture WG, Page 4
Moving Functionality from Core Network to RAN
Common RAN Control Plane (CP) framework
Services & Applications

› New more efficient common S1*


S1*
4) Mobility,
initial access (RACH and 5) D2D

Assign services to resources


LTE-NR tight communication
paging) integration 1) UE in IDLE or
INACTIVE mode,
camping
› Robust mobility to support LTE
all 5G scenarios, 1-100 Zzz
GHz NR z
X2*
X2*

› Tight integration with LTE NR


2) Initial access, UE
3) Mobility, enters CONNECTED
NR
› D2D communication multi-
connectivity, BF
state

integral part of the RAN mobility


design

› Agile RM Framework
METIS-II, Architecture WG, Page 5
RAN Protocol Stack Considerations
Static
Video Streaming Smart Grid
Temperature Sensor
(xMBB) (uMTC)
(mMTC)
There is consensus on the
State handling
fact that different services in HO measurements
State handling
RRC omitted
optim. for reduced optim. for reduced
5G will need tailored RAN/CN signaling state change latency

network functions in 5G (see Potential omitting of Potential omitting of


examples on right). PDCP ciphering and header default ciphering and header
compression compression

It is likely that this can be RLC


Unacknowledged
mode only
default
Acknowledged
mode only
achieved through
parameterization or (de-) H-ARQ optimized
H-ARQ omitted for
MAC default low-latency, RACH
selection of generic network for coverage
prioritization
functions Coding optimized Coding optimized Coding optimized
PHY for coverage, for very large for short payloads,
METIS-II, Architecture WG, Page 6
energy efficiency payloads low latency
Functionality on a Faster Time Scale
Dynamic AIV Reconfiguration
It is envisioned that software-defined
X MHz
a) networking approaches enable a fast
AIV AIV reconfiguration of air interface variants
AIV1
2 2 (AIVs) depending on traffic and service
f1 f2 MHz needs, new services rolled out etc.
Such reconfiguration would include
Y MHz
b) • Activation of new AIVs, with the
AIV AIV specific chaining of network
AIV1 functions needed
2 2
f1 f2 MHz • Change in key parameters of AIVs,
such as bandwidth, numerology
SDN-enabled etc.
dynamic reconfiguration of AIVs
Functionality on a Faster Time Scale
Dynamic AIV Reconfiguration
It is envisioned that software-defined
X MHz
a) networking approaches enable a fast
AIV AIV reconfiguration of air interface variants
AIV1
2 2 (AIVs) depending on traffic and service
f1 f2 MHz needs, new services rolled out etc.
Such reconfiguration would include
Y MHz
b) • Activation of new AIVs, with the
AIV AIV specific chaining of network
AIV1 functions needed
2 2
f1 f2 MHz • Change in key parameters of AIVs,
such as bandwidth, numerology
SDN-enabled etc.
dynamic reconfiguration of AIVs
Functionality on a Faster Time Scale
Agile Traffic Steering and Resource Management (1)

Core Network

METIS-II envisions the usage of a Service Flows


hierarchical traffic steering and High-Priority
Low-Priority Flows
Flows
resource management, with an AIV- Outer Access Network
Layer (AN-O) Radio link level
agnostic Access Network Outer layer feedback
(AN-O) receiving fast feedback from
AN-Inner
AIV-specific Inner Access Network NB1 NB2 NB3 (AN-I)
layers (AN-I)
Low-Reliability,
High-Reliability
High Capacity Link
Links
5G-UE
Agile Traffic Steering and Resource Mgmt
Functionality on a Faster Time Scale
Agile Traffic Steering and Resource Management (2)
Design Recommendation #1+:

Enable Dynamic Multi-AIV Resource Mapping


across novel & legacy AIVs avoiding hard
handovers.

RAN Design Implications

 Real-time Radio Link Feedback from AIVs.


 Low latency & High capacity X2*.

+: For more recommendations see the back-up slides


Functionality on a Faster Time Scale
Agile Traffic Steering and Resource Management (2)

AN-O

AN-I
Functionality on a Faster Time Scale
Agile Traffic Steering and Resource Management (2)

• Real-time SLA Monitoring to Guarantee Slice


Requirements
• Semi-dynamic Adaptation of Air Interface Variants
(AIVs) to adapt to network changes
AN-O

 The border between management & control


disappears
 From Requirements to SLA Fulfillment

AN-I
Functionality on a Faster Time Scale
Agile Traffic Steering and Resource Management (2)

Fast Fading Tracking via Real-time Radio Link Feedback


Slow UE: Speed 0.1 m/s Fast UE: Speed 10 m/s

AN-O

AN-I

LTE: 2 GHz (15 cm) & NR: 15 GHz (2 cm)


Functionality on a Faster Time Scale
Agile Traffic Steering and Resource Management (2)

Fast Fading Tracking


Performance of Dynamic
via Real-time
Multi-AIVRadio
Resource
Link Feedback
Mapping
Latency
Slow UE: Speed 0.1 m/s FastThroughput
UE: Speed 10 m/s

AN-O

AN-I

LTE: 2 GHz (15 cm) & NR: 15 GHz (2 cm)


An Evolved RAN Architecture
Possible UP/CP Split Constellations
S1* S1*-U S1*-C S1*-C
S1*
S1*-U
S1*
UP-H CP-H CP-H
CP-H
UP CP Split options M7, M8
UP CP
UP-L CP-L UP CP-L UP
Split options M1-M3 CP-L
RU RU RU RU RU
Stand-alone BTS Classical baseband pool Common split in UP and CP Higher CP layers CP (almost) fully
approach, possibly with and centralized handling of centralized (e.g. centralized (e.g. with
improved interface asynchronous functions, RRC), but classical centralized MAC
for good coordination stand-alone UP scheduler), imposing
Note that self-xhauling over a wireless interface while still relaxing latency operation higher latency
may rather suggest a stand-alone BTS case with requirements constraints on interfaces
minimized bandwidth and most relaxed latency
requirements
Network Slicing: RCM-specific Function
Automotive eHealth
Network Slicing: RCM-specific Function
Automotive eHealth

RACH for uMTC Group RACH for mMTC


Network Slicing: Inter-RCM Function
AN-O AN-I

AIV specific
Multi-Slice RM Gain
Core/Transport AaSE
scheduler

AIV specific context


Network slice information e.g. load
specific data flow indication, use case
support

Data flow with SLA


information

SLA monitoring,
AIV specific QoS mapping

AIV specific QoS


information

QoS KPIs

QoS to SLA mapping

SLA fulfilled or not


Way Forward (1/2)
Way Forward (2/2)
Thank You
http://www.metis2020.com

S-ar putea să vă placă și