Worlds Largest 4 G Network Feb 2106 BS

World’s Largest 4 G Network Feb, 2106 BS ~1. 1 M 4 G BSs ~720 K@2014 Terminal Sales volume 210 M+ Types 1000+, Price <50$ ~100 M@2014 Pg 2 | Subscriber ~360 M 4 G Subscribers ~90 M@2014 Coverage ~1. 2 B pop (~86%), Reach Villages ~330 Cities@2014

5 G Era: Rethink Fundamentals Green Communication Research Center established in Oct. 2011, initiated 5 G Key Tech R&D. Rethink Shannon To start a green journey of wireless systems Rethink Ring & Young For no more “cells” Rethink Signaling & Control To make network application/load aware Soft Rethink Antenna To make BS invisible Rethink Spectrum & Air Interface To enable wireless signal to “dress for the occasion ” Rethink Fronthaul Super Fast To enable Soft RAN via NGFI Rethink Protocol Stack To enable flexible configuration of diversified access points and optimal baseband function split between BBU pool & RRS “Towards Green & Soft: A 5 G Perspective” IEEE Comm. Mag, Vol. 52, Feb. 2014

5 G New Requirements Immersive Seamless Ultra Dense Tactile Massive Ultra Reliable Seamless Coverage, Hot Spot High Capacity, Low-power Massive-connection, Low-latency Ultra-reliable Pg 4 | New design principles, new key technologies, …

E 2 E 5 G System Key Technologies Green and Soft SDN/NFV Core Network Transportation network PTN PON UCN (enabled by C-RAN/NGFI) SDAI (enabled by MCD) RAN Low Freq. High Freq. New RIT Seamless wide-area coverage Hotspot & high data rate Massive-MTC RIT Mission-Critical RIT Low-power & massive-connections Low-latency & high-reliability

User Centric RAN (UCN): C-RAN a key enabler Service Awareness Autonomous Network CN-RAN Repartition Turbo Charged Edge Network Slice as a Service Localized data, service & forwarding Unified Access & Seamless RAN restructure Mobility Multi-Connectivity & Flexible Topology Pg 6 | Framework of radio access network

Fronthaul challenges surfaced with C-RAN (2011) C-RAN • • Centralized Control and/or Processing Collaborative Radio Real-Time Cloud Clean System Target CMCC has conducted extensive trials to seek costeffecitive FH solutions in LTE era Distributed BTS Traditional BTS Fronthaul used to be an issue for LTE C-RAN: CPRI Compression and WDM Pg 7 |

Traditional FH solutions must be revisited in 5 G era Pg 8 |

NGFI (x. Haul) Traffic dependent Support of statistical multiplexing Support of cell coordination Antenna independent Radio interface technological neutrality Pg 9 | • The key to achieve FH interface redesign lies in the function re-split b/w BBU and RRU • The new NGFI will further lead to re-design of underlined transport networks with packet switching capability

Decoupling Ant. /Non-ant. Related Processing 12000 The evaluation of the existing FH Interface (CPRI, OBSAI): ØThe FH bandwidth is proportional to the number of antennas. ØThe FH bandwidth is at least 2 orders of magnitude higher than BH bandwidth. Considering the C-RAN centralized deployment and the technologies of 5 G, FH is facing a bandwidth explosion. The number of Carriers in C-RAN The number of antennas FH BW based on the existing FH Interface 100 8 1 Tbps 100 128 16 Tbps Pg 10 | 10000 8000 Radio BW 6000 FH BW BH BW 4000 2000 400 30 0 0. 2 0. 013 GSM 1. 6 0. 384 TD-SCDMA 20 150 8 -antenna TD-LTE Decoupling antenna/non-antenna related processing: ØIt is proposed that antenna related functions should be moved from the BBU to the RRH. ØFH bandwidth will decrease significantly if the BBU/RRH function split can decouple the non-antenna related processing and the antenna related processing.

Decoupling cell/UE Processing The existing FH interface is a constant bit rate interface, which is load independent and does not match with the features of mobile traffic. Decoupling cell/UE processing: Pg 11 | ØCell processing is irrelevant to traffic load and is fixed no matter how many UEs are active. ØFH bandwidth will be lower and load dependent. ØIn C-RAN Mode, cell/UE processing decoupling can further help reduce power consumption. ü When the traffic load is low, part of C-RAN cloud resource can be shut down. ü When there is no active UE, BBU software can be switched to a dormant state

Decoupling UL/DL Processing Some symmetrical BBU-RRU function re-split solutions （Taking LTE as an example） There are many different aspects between UL and DL: ØGenerally, the DL rate is not equal to the UL rate. ØThe bandwidth of UL and DL are not always in the same order of magnitude. Ø The bit width of UL is usually larger than that of the DL. Decoupling UL/DL processing Ø NGFI design should take into account the asymmetrical function split solutions. Ø For example, the function split solution 3 could be used for UL while the solution 4 used for DL. Pg 12 |

NGFI Progress • Lead of IEEE 1914 (NGFI) WG in IEEE • 1 st NGFI WS held & NGFI WP released in June, 2015 下一代前传网络接口 白 皮 书 - Mo. U signing with Broadcom, Intel, Alcatel-Lucent, Hua. Wei, ZTE, Nokia, Xilinx & Altera • Co-founder of IEEE 1904. 3 White Paper of Next Generation Fronthaul Interface • Leading the project of NGFI in CCSA • NGFI as the key component in NGMN 5 G WP, Fu. TURE 5 G WP • NGFI/FH promotion and study in ITU-T, IEEE and 3 GPP • NGFI Paper in IEEE Communication Magazine & GLOBECOM 2015 “Rethink Fronthaul for Soft RAN”, IEEE Comm. Mag. 2015 “NGFI, The x. Haul”, GLOBECOM 2015 Version 1. 0 June 4 th, 2015 China Mobile Research Institute Alcatel-Lucent Nokia Networks ZTE Corporation Broadcom Corporation Intel China Research Center Pg 13 | NGFI feasibility study with Xilinx White Paper on NGFI released in June 2015

Challenges Ahead Mapping of FH to Ethernet packet, undergoing in IEEE 1904 WG Stage 1: CPRI encapsulation (applicable to 2 G, 3 G and 4 G) Stage 2: NGFI encapsulation for 5 G Latency To meet RT requirement of wireless communication Limited latency budget for FH (e. g. 250 us defined in NGMN for LTE C-RAN) Synchronization LTE Frequency sync. of 0. 05 ppm & phase accuracy of +/-1. 5 us Higher in 5 G Transport of FH packets Over IP, MPLS, etc. E 2 E Qo. S O&M, protection etc. Pg 14 |

IEEE NGFI (1914) WG • • Pg 15 | Target: efficient & scalable FH for 5 G Officially approved: Feb. 2016 Sponsor: IEEE COM/SDB 7 founding companies with more than 50 subscribers so far from ~30 companies Scope of 1914. 1 project: - NGFI transport network architecture - Requirements - Function split analysis for LTE 7 Founding members http: //grouper. ieee. org/groups/1914/ The first NGFI WG meeting, April 25 -28, San Jose, CA Contact: huangjinri@chinamobile. com

Thank you For more information, please contact icl@chinamobile. com Pg 16 |