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 (bomin.li@comcores.com)

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