5 g fundamental

1. www.cubeoptics.com
3GPP
TR 38.801
eCPRI v1 E
IID
ID
D
C
B
A
Small Cell Forum
NGMN
MAC-PHY (nFAPI)
Split MAC
RLC-MAC
PDCP-RLC
MAC-PHY
II
I II
7-2x (O-RAN)
III
III
IIIb
~IV
8
7-1
7-2
7-3 (DL only)
6
5
4
3
2
1
High Layer Split Low Layer Split (6 & 7)
Split
Naming
Overview
Split 1 2 3 4 5 6 7-3 7-2 7-2x 7-1 8
Pros
+ Low bandwidth requirements.
+ Bitrate scales with MIMO layers.
+ Separate User Plane and centralized
RRC/RRM.*
+ It may in some circumstances provide
benefits in handling some edge comput-
ing or low latency use cases where the
user data needs to be located close to
the transmission point.*
+ Fundamentals for achieving a PD-
CP-RLC split have already been stan-
dardized for LTE Dual Connectivity.*
+ The 2-2 option enables centralization of
the PDCP layer.*
+ Option 2-2 allows a separate UP and a
centralized RRC/RRM.*
+ Very Low bandwidth requirements.
+ Low latency requirements.
+ More robust under non-ideal transport
conditions.*
+ Possibility of reduced processing and
buffer requirements in DU.*
+ In option 3-2 Rx RLC is placed in the
CU, there is no additional transmission
delay of PDCP/RLC reestablishment
procedures.*
+ Low bandwidth requirements.
+ Bitrate scales with MIMO layers.
+ Low bandwidth requirements.
+ Reduced latency requirements if HARQ
processing and cell-specific MAC func-
tionalities are performed in DU.*
+ Efficient interference management
across multiple cells and enhanced
scheduling technologies such as CoMP,
CA, etc.*
+ Bitrate scales with MIMO layers
+ Significant bandwidth reduction com-
pared to split option 7-3.
+ Joint Transmission is possible.*
+ Centralized scheduling is possible.*
+ Allows resource pooling for layers in-
cluding and above MAC.*
+ Bitrate scales with MIMO layers
+ Reduced bandwidth requirements com-
pared to split option 7-1.
+ Coordinated multi-point schemes are
possible if CU/DU are colocated.*
+ Transmit and receive joint processing is
possible.*
+ Bitrate scales with MIMO layers
+ Reduced bandwidth requirements com-
pared to split option 7-1.
+ Coordinated multi-point schemes are
possible if CU/DU are colocated.*
+ Transmit and receive joint processing is
possible.*
+ Simplified interface
+ Open interface protocol specifically
designed to enable interoperability
between RUs and DUs from different
vendors.
+ Bitrate scales with MIMO layers
+ Reduced bandwidth requirements com-
pared to split option 7-1.
+ The required bitrate is more than half of
split option 8.
+ Coordinated multi-point schemes are
possible if CU/DU are colocated.*
+ Transmit and receive joint processing is
possible.*
+ Small and cost effective RU.
+ Easy to centralize CU/DU enabling co-
ordinated multi-point (CoMP) schemes.*
+ Majority of processing can be central-
ized at a BBU hotel or CU-pool.*
+ RUs can be used for different genera-
tions of RAT (GSM, 3G, 4G)
Cons
- Very complex and expensive DU/RU.
- It‘s not clear if this option can support
aggregation based on alternative 3C.*
- Coordination of security configurations
between different PDCP instances for
Option 2-2 required.*
- Split 3-1 is more latency sensitive than
3-2 due to the ARQ in CU and not DU.*
- No benefits for LTE.* - Complex interface between CU and
DU.*
- Difficulty in defining scheduling opera-
tions over CU and DU.*
- Limitations for some CoMP schemes.*
- May require subframe-level timing inter-
actions between MAC layer in CU and
PHY layers in DUs.*
- Round trip fronthaul delay may affect
HARQ timing and scheduling.*
- High bandwidth requirements.
- Relatively high latency requirements
- Complex timing for RU and CU/DU link.*
- High bandwidth requirements.
- Relatively high latency requirements
- Complex timing for RU and CU/DU link.*
- High bandwidth requirements.
- Relatively high latency requirements.
- Still relatively high bandwidth
requirement especially for the uplink.
- Bandwidth scales with number of RUs.*
- Very latency constrained.
- Complex timing for RU and CU/DU link.*
- Highest bandwidth requirements of all
functional split options.
- Bandwidth scales with number of RUs.*
- Very latency and jitter constrained.
- Distance between RU and DU/CU limit-
ed to ~20km due to latency constraint.
- Interoperability between radio equip-
ment vendors not specified
Use
Cases
> Best suited for low latency and/or edge
computing scenarios.
> Suited for high layer split between CU
and DU. Very latency tolerant enabling
distances up to 40km.
> Low bitrate and latency insensitive
midhaul connections between CU and
DU with non-ideal transport conditions.*
> No specific advantage for use cases. > Ideal for scenarios where distances
greater than 20km between DU and CU
need to be bridged.
> Ideal for small cell deployments. > Suited for setup with limited fiber capac-
ity in the fronthaul.
> Current 5G eCPRI radios use this split
option.
> Ideally suited for virtualized RAN and
virtual DU running on general purpose
processing platforms.
> High fiber capacity available between
radio and centralized location.
> High fiber capacity available between
radio and centralised location.
> Real time communication applications.
> Possible to integrate in Ethernet based
networks using Radio over Ethernet.
Double
Split
High Layer Split
Low
Layer
Split
CU
and
DU
Mapping
CU/DU RU
CU DU/RU
CU/DU RU
CU/DU RU
Midhaul
Fronthaul
Fronthaul
Fronthaul
Backhaul
Backhaul
Backhaul
Backhaul
Option 8
CPRI
Option 2
Option 6
nFAPI
Option 7
eCPRI IID
DU RU
Fronthaul
Backhaul CU Midhaul Option 2 & 6
DU RU
Fronthaul
Backhaul CU Midhaul Option 2 & 7.2
Core
Network
CU/DU RU
Fronthaul
Backhaul
Option 7-2x
O-RAN
High Layer Split
+ Drastically reduced Bandwidth
+ Ideal for non-mobile = FWA
+ Latency Tolerant = long distances
+ Processing in RRH = URLLC
- CoMP extremely complex or even impossible
- Complex and expensive RRH (size, heat, cost)
Double Split
+ CU can easily be virtualized
+ Optimal for mobile and URLLC
+ Reduced cost
+ Good scalability
- High bandwidth and latency fronthaul
requirements
Low Layer Split
+ Ideal for CoMP = mobile
+ Cost effective RRH
- High Bandwidth
- Bandwidth scales with antenna ports (8, 7-1)
- Very tight latency requirements
Based on 100MHz, 32 Antennas
UL DL
157.3 157.3
UL DL
60.4 9.2
UL DL
15.2 9.8
UL DL
15.2 9.8
UL DL
15.2 9.8
UL DL
7.1 5.6
UL DL
7.1 5.6
UL DL
4.5 5.2
UL DL
3 4
UL DL
3 4
UL DL
3 4
Based on 100MHz, 8 layers, 256 QAM
Band-
width
(LTE
/
Gpbs)
Scala­
bility
n.a. Scales with MIMO layers Scales with antenna port
HARQ Loop - Very tight latency requirements
Latency and Jitter
constrained
Lat.-
Req.
250µs
10ms ~100µs
Latency tolerant 1.5 - 10ms
Retrans-
mission
buffer
Header
com­
pression
Num­
bering
Ciphering
Add
PDCP
header
Trans­
mission
buffer
Add RLC
header
Multi­
plexing
HARQ
DL-SCH
data
transfer
CRC
attach
Coding +
block seg
Rate
matching
Scram-
bling
Modula-
tion
Layer
mapper
Pre-
coding
Resource
Element
Mapper
Beamf.
Port exp.
iFFT
Cyclic
Prefix
Insertion
Analog
Conver-
sion
RLC
control
Antenna
ports
Data
Segmen-
tation
Retrans-
mission
buffer
Controller Scheduler
Functional
View
The lower the split option, the less
functions are in the CU.
The higher the split option, the
less functions are in the RU.
Downlink
Transport
Blocks
Coded
Block
Code-
words
Code-
words
Layer
N
symbols
Sub­
carriers
IQ
Symbols
Antenna
ports
PDCP
SDU
RLC
SDU
MAC
SDU
Transport
Blocks
Symbols
Antenna
N
symbols
Sub­
carriers
IQ
Symbols
Logical Link Control (LLC) - OSI Layer 2(b) MAC - OSI Layer 2(a)
OSI Layer 2 - Data Link OSI Layer 1 - Physical
RF
Low PHY
High PHY
Low MAC
High
MAC
Low RLC
High RLC
PDCP
IP
Protocol
Stack
5G Fundamentals : Functional Split Overview
5G Fundamentals Functional Split Overview
* 3GPP TR 38.801 V14.0.0 (2017-03): „Study on new radio access technology: Radio access architecture and interfaces.”
Rev.
3
December
2019