Fronthaul architecture towards 5G -...
Transcript of Fronthaul architecture towards 5G -...
Aleksandra Checko, MTI
8/22-24/2016
In collaboration with:
MTI Radiocomp: Andrijana Popovska
Avramova, Morten Høgdal, Georgios
Kardaras
DTU: Michael Berger, Henrik L. Christiansen
EU project HARP consortium
Fronthaul architecture towards 5G
Multiplexing gains analysis
Challenges/solutions for fronthaul network
Compliance with IEEE Standards Policies and Procedures
Subclause 5.2.1 of the IEEE-SA Standards Board Bylaws states, "While participating in IEEE standards development activities, all participants...shall act in accordance with all applicable laws (nation-based and international), the IEEE Code of Ethics, and with IEEE Standards policies and procedures."
The contributor acknowledges and accepts that this contribution is subject to
• The IEEE Standards copyright policy as stated in the IEEE-SA Standards Board Bylaws, section 7, http://standards.ieee.org/develop/policies/bylaws/sect6-7.html#7, and the IEEE-SA Standards Board Operations Manual, section 6.1, http://standards.ieee.org/develop/policies/opman/sect6.html
• The IEEE Standards patent policy as stated in the IEEE-SA Standards Board Bylaws, section 6, http://standards.ieee.org/guides/bylaws/sect6-7.html#6, and the IEEE-SA Standards Board Operations Manual, section 6.3, http://standards.ieee.org/develop/policies/opman/sect6.html
2
Fronthaul architecture towards 5G
Multiplexing gains analysis
Challenges/solutions for fronthaul network
Date: 2016-08-23
Author(s):
Name Affiliation Phone [optional] Email [optional]
Aleksandra CheckoMTI (Microelectronics
Technology, Inc.)
aleksandra.checko
@mtigroup.com
IEEE 1914.1 TFNGFI
Bomin Li ([email protected])
5G requirements and challenges
Source: netmanias
To carry 20Gb/s as I&Q signals (64-QAM ¾) over CPRI will require a CPRI data rate of at least 325Gb/s
Not acceptable for NGFI
New functional split is needed
– User processing in BBU
– Cell processing in RRH
– Split functions, still benefit from C-RAN (e.g. CoMP)
Variable bit rates on fronthaul
Agenda for today
– How big are the multiplexing gains in C-RAN?
– How to optimize fronthaul network?
Studying multiplexing gains
- towards quantifying benefits of C-RAN
Multiplexing and pooling gain defined
Exploring the tidal effect
Exploring different application mixed and measurement methods
Cloud
RRH 1
RRH 2
RRH n
...
BBU Pool
Mobile Backhaul Network
Aggregated Traffic (h)
24 h
Cloud-RAN
∑ < n
RRH 1
RRH 2
RRH n
...
Mobile Backhaul Network
BBU 2
BBU nBBU 1
RAN with RRHs
∑ = n
Multiplexing gains are available for any shared resources
0
20
40
60
0 6 12 18 24
Load
Time (h)
Data from China Mobile [1]Office Residential
How big are the multiplexing gains in C-RAN?
Cloud
RRH 1
RRH 2
RRH n
...
BBU Pool
Mobile Backhaul Network
Aggregated Traffic (h)
24 h
Cloud-RAN
∑ < n
RRH 1
RRH 2
RRH n
...
Mobile Backhaul Network
BBU 2
BBU nBBU 1
RAN with RRHs
∑ = n
𝑀𝐺𝑝𝑟𝑖𝑛𝑡𝑒𝑟𝑠 =#𝑝𝑒𝑜𝑝𝑙𝑒
#𝑝𝑟𝑖𝑛𝑡𝑒𝑟𝑠
Multiplexing gains are available for any shared resources
Variable bit rate on the fronthaul network
– Bursty traffic
BBUs in the pool – Pooling Gain (PG)
– Processing resources
– Power
1
1
1
1
RRH 1
RRH 2
RRH n
...
Mobile Backhaul Network
BBU 2
BBU nBBU 1
RAN with RRHs
∑ = n
RRH 1
RRH 2
RRH n
...
Mobile Backhaul Network
BBU 2
BBU nBBU 1
RAN with RRHs
∑ = n
RRH 1
RRH 2
RRH n
...
BBU Pool
Mobile Backhaul Network
Aggregated Traffic (h)
24 h
Cloud-RAN
∑ < nCell-processing
User-processing
Multiplexing gains in C-RAN
Sources
– Aggregation of bursty traffic
– Tidal effect
– Different functional splits
• On BBU/fronthaul
Impact energy and cost savings
– On user-data dependent resources
– Processing: ctrl + cell + user
• 3-12% on downlink, 17-33% on uplink of total baseband processing [2]
– Power: ctrl + cell + user
• 2-24 % of total base station consumption [2]
MG
MG
Bit-level
processing
MAC
RLC
PDCP
CP
Resource mapping
IFFT
QAMAntenna mapping
Use
r p
roce
ssin
g
Variable
bit rate
Constant
bit rate
PGPDCP
PGRLC
PGMAC
PGBLP
PGQAM
PGRM_IFFT
PGCP
P
H
Y
Ce
ll
pro
ce
ssin
g
Exploring tidal effect – analytical approach
0.8
0.9
1.0
1.1
1.2
1.3
1.4
1.5
1.6
1.7
0 10 20 30 40 50 60 70 80 90 100M
ult
iple
xin
g ga
inOffice cells (%)
Multiplexing gain for different locations, for different mix of office and residential
cellsLondon/LosAngeles
New York
Hong Kong
China Mobile
China Mobile,simulated
0
0.5
1
1.5
2
2.5
0 6 12 18 24
DL
dat
a
Time (h)
Traffic in London, from MIT/Ericsson [3]
City of London Newham
0
0.5
1
1.5
2
2.5
0 6 12 18 24
DL
dat
a
Time (h)
Traffic in New York, from MIT/Ericsson [3]
Lower Manhattan Ridgewood
[4]
Exploring application mix – OPNET simulations [5]
Results for 30% office, 70% residential cells, web and video traffic
Traffic burstiness contributes to multiplexing gain up to 6
Switch
….
3 office cells
….
7 residential
cells
MAC
IP
TCP
Application
MAC
IP
TCP
Application
MAC MAC
Traffic (h)
24 h
Traffic (h)
24 h
Aggregated
Traffic (h)
24 h
BBU Pool
0
1
2
3
4
5
6
7
0 10 20 30 40 50 60 70 80 90 100
Mu
ltip
lexin
g g
ain
Web traffic (%)
100 ms
1 s
Discussion
• PG from analyzing compute resource utilization, in Giga Operations Per Second (GOPS): 1.09 - 1.37, source: Werthmann et al. [6]. Tidal effect not accounted.
• PG based on population in different districts in Tokyo: 4, source: Nambaet al. [7]
• MG from tidal effect: 1-1.3
• MG from traffic burstiness: up to 6
• Fraction of it impacts baseband resources (3-33%) to achieve PG
MG up to 6 achievable on fronthaul, fraction of it achievable on BBU side
New functional split should result in bursty traffic being as “low” as possible (closest to BB-RF – traditional RRH BBU split) to benefit from C-RAN (e.g. spectral efficiency)
Fronthaul transport network
- Towards NGFI
Transport options
Synchronization challenge: application of IEEE 1588
Delay challenge: application of TSN
Possible transport solutions
Technology is ready
(capacity not necessarily)
CPRI / OBSAI
CPRI / OBSAI
WDM
CPRI / OBSAI
CPRI / OBSAI
Microvawe
Compression Compression
Eth/CPRI/ OBSAI
Eth/CPRI/ OBSAI
Ethernet
CPRI/ OBSAI
CPRI/ OBSAI
OTN
BBU Pool
RRH
CPRI / OBSAI Point to point
WDM-PON
Widely deployed (reuse!)
Dedicated links
Shared links
Aggregation
Multiplexing gains on BBU and links
Switching
! Fronthaul cost savings vs problems with delays and synchronization
! Synchronous CPRI vs asynchronous Ethernet
! Data delay: 100-400 us, ≈constant
BBU Pool
GPS/1588 Slave
RRH
RRH
Sync delivered together with data (CPRI)
Ethernet network
Eth switches
BBU Pool
1588 Master
1588 Slave
RRH
CPRI2Ethgateway
RRH
CPRI2Ethgateway
1588 Slave
Shared Ethernet for cost-saving and flexibility [8]
Timing in fronthaul
Timing is really important
Frequency of transmission
Handover, coding, cooperative techniques, positioning
00
0102
00
01
02
Device A
Device B
Current solutions for timing distribution GPS
PHY layer clock – SyncEth
Packet-based timing
IEEE 1588v2 (PTP)
Multiple
Requirements (4G) Frequency error LTE – A TDD/FDD: ±50
ppb
Phase error LTE-A with eICIC/CoMP: ± 1.5 - 5 μs, MIMO: 65 ns, positioning: ± 30 ns
What are the requirements for RoE and 5G?
How to reduce queueing delays?
Preemption (switch upgrade required)
Scheduling and source scheduling
A
BB A
A
BB A
No scheduling – B delay non deterministic
Scheduling – B delayed initially, delay deterministic
BBU pool BBU pool
A
B
A
AB B
B
Input (high priority)
Input (low priority)
Output (no preemption)
Output (with preemption)
Preemption
A
B
AB
AB B
IEEE 802.1, Time Sensitive Networking task force
Frame preemption (802.1Qbu)
Scheduled traffic (802.1Qbv)
Time-Sensitive Networking for Fronthaul (profile definition, 802.1CM)
Exemplary architecture
With control solution – e.g. SDN
With synchronization solution – e.g. IEEE 1588
With delay minimization solution – e.g. TSN
SDN
Ethernet network
Eth switches
BBU Pool
1588 Master
1588 Slave,CPRI2EthGateway,
source scheduling
RoE RRH, 1588 Slave, Source scheduling
Legacy RRH
Conclusions and final remarks
Costs
– 2x2 MIMO, 20 MHz LTE,
15+1 CPRI 2.5 Gbps
– 3 sectors? 7.5 Gbps
– Tens of BS over long
distance? 100+ Gbps
Costs vs savings
C-RAN benefits?Fronthaul
cost?
Savings
– Equipment
– Energy
Benefits from cooperative
techniques
Re-definefronthaul
New function split
Ethernet-based FH
Cost savings
Joint design, joint benefits
Conclusions, proposals to 1914
Optimal functional split is needed to reduce data rate and benefit from
multiplexing gains on fronthaul, while exploiting benefits of C-RAN
One split probably won’t fit all – possible reconfiguration options are
interesting
Multiplexing gains are possible on BBU resources (on 3-33% of
resources), and for variable bit rate split also on fronthaul.
Industry shows a strong interest in packet-based fronthaul.
Ethernet-based fronthaul with traffic scheduling and/or preemption has
the potential to meet mobile networks’ requirements while being cost-
efficient.
Thank you for your attention
References
[1] C-RAN The Road Towards Green RAN. Tech. rep. China Mobile Research Institute, October 2011
[2] C. Desset, et al. “Flexible power modeling of LTE base stations”. In: Wireless Communications and Networking Conference (WCNC), 2012 IEEE, Apr. 2012, pp. 2858–2862.
[3] Many cities. MIT Senseable City Lab. [cited: January 2016]. URL: http://www.manycities.org/
[4] A. Checko, H. Holm, and H. Christiansen. “Optimizing small cell deployment by the use of C-RANs”. In: European Wireless 2014; 20th European Wireless Conference; Proceedings of. 2014 VDE
[5] A. Checko1st, A. P. Avramova1st, H. L. Christiansen, and M. S. Berger. “Evaluating C-RAN fronthaul functional splits in terms of network level energy and cost savings”. in Journal of Communications and Networks, vol. 18, no. 2, pp. 162-172, April 2016.
[6] T.Werthmann, H. Grob-Lipski, and M. Proebster. “Multiplexing gains achieved in pools of baseband computation units in 4G cellular networks”. In: Personal Indoor and Mobile Radio Communications (PIMRC), 2013 IEEE 24th International Symposium on. Sept. 2013, pp. 3328–3333.
[7] S. Namba, et al. “Colony-RAN architecture for future cellular network”. In: Future Network Mobile Summit (FutureNetw), 2012. July 2012, pp. 1–8
[8] A. Checko, A. Juul, H. Christiansen, M. S. Berger, "Synchronization Challenges in Packet-based Cloud-RAN Fronthaul for Mobile Networks“, IEEE ICC 2015
A. Checko, H. Christiansen, Y. Yan, L. Scolari, G. Kardaras, M.S. Berger and L. Dittmann „Cloud RAN for Mobile Networks - a Technology Overview”, IEEE Communications Surveys & Tutorials