See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/316657605

Cloud Mobile Networks

Book · January 2017

DOI: 10.1007/978-3-319-54496-0

CITATIONS READS

22 1,019

2 authors, including:

Mojtaba Vaezi
Villanova University
119 PUBLICATIONS 2,718 CITATIONS

SEE PROFILE

All content following this page was uploaded by Mojtaba Vaezi on 16 November 2019.

The user has requested enhancement of the downloaded file.

Wireless Networks

Mojtaba Vaezi
Ying Zhang

Cloud Mobile
Networks
From RAN to EPC
Wireless Networks

Series editor
Xuemin (Sherman) Shen
University of Waterloo, Waterloo, Ontario, Canada
More information about this series at http://www.springer.com/series/14180
Mojtaba Vaezi Ying Zhang
•

Cloud Mobile Networks

From RAN to EPC

123
Mojtaba Vaezi Ying Zhang
Princeton University Hewlett Packard Labs
Princeton, NJ Fremont, CA
USA USA

ISSN 2366-1186 ISSN 2366-1445 (electronic)

Wireless Networks
ISBN 978-3-319-54495-3 ISBN 978-3-319-54496-0 (eBook)
DOI 10.1007/978-3-319-54496-0
Library of Congress Control Number: 2017934618

© Springer International Publishing AG 2017

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt from
the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, express or implied, with respect to the material contained herein or
for any errors or omissions that may have been made. The publisher remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.

Printed on acid-free paper

This Springer imprint is published by Springer Nature

The registered company is Springer International Publishing AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Preface

In view of cost-effective implementations, scalable infrastructure and elastic

capacity on demand, virtualization and cloud computing are now becoming the
cornerstones of any successful IT strategy. Emerging cloud computing technolo-
gies, such as the software-defined networking and Internet of things, have been
thoroughly investigated for data computing networks, while less attention has been
paid to radio access network virtualization, be it hardware or software elements.
Today, the focus of research in wireless and cellular networks has shifted to
virtualization and cloud technologies, so that incorporation of cloud technologies,
network functions virtualization, and software-defined networking is essential part
in the development process of 5G cellular communications system, expected to be
commercialized by 2020. These technologies are expected to affect different parts of
cellular networks including the core network and radio access network (RAN).
Cloud RAN has emerged as a revolutionary approach to implementation,
management, and performance improvement of next-generation cellular networks.
Combined with other technologies, such as small cells, it provides a promising
direction for the zettabyte Internet era. The virtualization of RAN elements is
stressing the wireless networks and protocols, especially when the large-scale
cooperative signal processing and networking, including signal processing in the
physical layer, scheduling and resources allocation in the medium access control
layer, and radio resources managements in the network layer, are centralized and
cloud computed.
The main motivation for offering this book stems from the observation that, at
present there is no comprehensive source of information about cloud RAN and its
interplay with other emerging technologies for network automation, such as the
software-defined networking, network functions virtualization, and wireless

v
vi Preface

virtualization. In addition to providing the latest advances in this area, we also

include research potentials and market trend in this field. We believe that it is
valuable to bring basic concepts and practical implementation of several related
areas together, to facilitate a better understanding of the entire area.

Princeton, USA Mojtaba Vaezi

Fremont, USA Ying Zhang
Contents

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Challenges of Today’s RAN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.1.1 Cost of Ownership . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.1.2 Capacity Expansion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.1.3 Energy Consumption and Carbon Emissions . . . . . . . . . . . 5
1.2 Cloud RAN - What Is the Big Idea? . . . . . . . . . . . . . . . . . . . . . . . 6
1.2.1 Advantages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.2.2 Challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.3 Related Technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.3.1 Network Virtualization . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.3.2 Network Functions Virtualization . . . . . . . . . . . . . . . . . . . . 9
1.3.3 Software-Defined Networking. . . . . . . . . . . . . . . . . . . . . . . 9
1.4 Outline of Chapters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2 Virtualization and Cloud Computing . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.1 What Is Virtualization? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.2 Why Virtualization? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.3 Network Virtualization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.3.1 Overlay Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.3.2 Virtual Private Networks . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.3.3 Virtual Sharing Networks . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.3.4 Relation Between Virtual Networks . . . . . . . . . . . . . . . . . . 19
2.4 Network Functions Virtualization . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.4.1 What to Virtualize? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.5 Wireless Virtualization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.5.1 State of the Art in Wireless Virtualization . . . . . . . . . . . . . 25
2.6 Cloud Computing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.6.1 Cloud Services Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.6.2 Types of Clouds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
2.6.3 Virtualization Versus Cloud Computing . . . . . . . . . . . . . . . 30
2.7 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

vii
viii Contents

3 Software-Defined Networks Principles and Use Case

Scenarios . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
3.2 SDN Drivers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
3.2.1 Technology Drivers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
3.2.2 Business Drivers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.3 Architecture and Principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.3.1 Split Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
3.3.2 Open-APIs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
3.3.3 Virtualization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
3.3.4 Control Plane Signaling Protocol . . . . . . . . . . . . . . . . . . . . 37
3.4 Use Case Scenarios . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
3.4.1 SDN in Mobile Backhaul Networks . . . . . . . . . . . . . . . . . . 37
3.4.2 SDN in the Cloud . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.4.3 SDN in NFV and Service Chaining . . . . . . . . . . . . . . . . . . 40
3.4.4 SDN Based Mobile Networks . . . . . . . . . . . . . . . . . . . . . . 41
3.5 SDN Design Implementation Considerations . . . . . . . . . . . . . . . . . 42
3.5.1 Scalability and High Availability . . . . . . . . . . . . . . . . . . . . 42
3.5.2 Reliability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
3.5.3 A Special Study: Controller to Switch Connectivity . . . . . . 44
3.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
4 Virtualizing the Network Services: Network Function
Virtualization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
4.1 NFV Architecture. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
4.2 NFV Use Cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
4.3 NFV Challenges. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
4.4 Service Function Chaining . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
4.4.1 Openflow-Based SFC Solution . . . . . . . . . . . . . . . . . . . . . . 52
4.4.2 Optical SFC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
4.4.3 Virtualized Network Function Placement . . . . . . . . . . . . . . 55
4.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
5 SDN/NFV Telco Case Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
5.1 Packet Core . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
5.1.1 Existing Solutions Problems . . . . . . . . . . . . . . . . . . . . . . . . 58
5.1.2 Virtualization and Cloud Assisted PC. . . . . . . . . . . . . . . . . 59
5.2 Virtualized Customer Premises Equipment. . . . . . . . . . . . . . . . . . . 62
5.2.1 Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
5.2.2 Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
5.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
6 Radio Access Network Evolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
6.1 Mobile Network Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
6.1.1 End-to-End Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
6.1.2 Packet Flow in a Mobile Network . . . . . . . . . . . . . . . . . . . 68
Contents ix

6.2 Base Station Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69

6.2.1 Legacy BS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
6.2.2 Architecture of Distributed BS . . . . . . . . . . . . . . . . . . . . . . 70
6.3 Mobile Fronthaul and Backhaul . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
6.3.1 Backhaul . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
6.3.2 Fronthaul . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
6.4 Trends in Wireless Network Densification . . . . . . . . . . . . . . . . . . . 76
6.4.1 Increasing Spectral Efficiency . . . . . . . . . . . . . . . . . . . . . . . 77
6.4.2 Interference Mitigation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
6.4.3 Millimeter Wave Cellular Systems . . . . . . . . . . . . . . . . . . . 79
6.5 Small Cells and HetNets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
6.5.1 Public Wi-Fi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
6.6 Distributed Antenna Systems (DAS) . . . . . . . . . . . . . . . . . . . . . . . 82
6.6.1 DAS or Small Cells? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
6.7 Network Sharing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
6.7.1 Challenges for eUTRAN Sharing . . . . . . . . . . . . . . . . . . . . 84
6.7.2 Standards Specifications . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
6.8 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
7 Cloud RAN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
7.1 Cloud RAN: Definition and Concept . . . . . . . . . . . . . . . . . . . . . . . 87
7.1.1 Centralization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
7.1.2 Virtualization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
7.1.3 Statistical Multiplexing . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
7.1.4 Multi-cell Cooperation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
7.2 Potentials and Benefits of Cloud RAN Architecture. . . . . . . . . . . . 93
7.2.1 Potentials of Cloud RAN . . . . . . . . . . . . . . . . . . . . . . . . . . 93
7.2.2 Benefits of Cloud RAN . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
7.3 Cloud RAN Implementation Scenarios. . . . . . . . . . . . . . . . . . . . . . 95
7.3.1 RAN Protocol Architecture. . . . . . . . . . . . . . . . . . . . . . . . . 95
7.3.2 Proposed Architectures for C-RAN . . . . . . . . . . . . . . . . . . 97
7.3.3 RAN as a Service (RANaaS) . . . . . . . . . . . . . . . . . . . . . . . 100
7.4 Cloud RAN or Small Cells? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
7.5 Cloud RAN Challenges and Research Frontiers . . . . . . . . . . . . . . . 101
7.5.1 Challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
7.5.2 Research Frontiers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
7.5.3 Edge Cloud . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
7.6 Cloud RAN Market . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
7.7 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
Abbreviations and Acronyms

3GPP 3rd Generation Partnership Project

AAS Active Antenna System
AIS Alarm Indication Signal
API Application Programming Interface
ARPU Average Revenue Per User
ARQ Automatic Repeat reQuest
ASIC Application Specific Integrated Circuit
AWGN Additive White Gaussian Noise
BBU Baseband Unit
BRAS Broadband Remote Access Server
BS Base Station
BSS Business Support System
BTS Base Transceiver Station
CAGR Compound Annual Growth Rate
CapEx Capital Expenditure
CB Coordinated Beamforming
CO2 Carbon dioxide
CoMP Coordinated MultiPoint
CPRI Common Public Radio Interface
CPU Central Processing Unit
C-RAN Cloud RAN
CS Coordinated Scheduling
CSI Channel State Information
DAS Distributed Antenna Systems
DC Data Center
DPI Deep Packet Inspection
D-RoF Digital Radio over Fiber
DSL Digital Subscriber Line
EDGE Enhanced Data rates for GSM Evolution
eNB Evolved NodeB

xi
xii Abbreviations and Acronyms

eNodeB E-UTRAN NodeB

EPC Evolved Packet Core
ETSI European Telecommunications Standards Institute
eUTRAN Evolved UTRAN
EVM Error Vector Magnitude
FDMA Frequency Division Multiple Access
GGSN Gateway GPRS Support Node
GPP General Purpose Processor
GPRS General Packet Radio Service
GSM Global System for Mobile Communications
GTP GPRS Tunneling Protocol
GWCN Gateway Core Network
HARQ Hybrid Automatic Repeat reQuest
HetNet Heterogeneous Network
HSPA High Speed Packet Access
HTTP Hypertext Transfer Protocol
IaaS Infrastructure as a Service
ICI Inter-Cell Interference
ICT Information and Communications Technology
IDS Intrusion Detection Systems
IMS IP Multimedia Subsystem
InP Infrastructure Provider
IoT Internet of Things
IP Internet Protocol
I/O Input/Output
ISP Internet Service Provider
ISSU In-Service Software Upgrades
IT Information Technology
JP Joint Processing
JT Joint Transmission
L1 Layer 1
L1VPN Layer 1 VPN
L2TP Layer 2 Tunneling Protocol
L2VPN Layer 2 VPN
L3VPN Layer 3 VPN
LAN Local Area Networks
LTE Long-Term Evolution
LTE-A LTE-Advanced
M2M Machine-to-Machine
MAC Media Access Control
MANO Management and Network Orchestration
MIMO Multiple Input Multiple Output
MIP Mixed Integer Programming
MME Mobility Management Entity
mmW Millimeter Wave
Abbreviations and Acronyms xiii

MOCN Multi-Operator Core Network

MSC Mobile services Switching Center
MU-MIMO Multi-User MIMO
MVNO Mobile Virtual Network Operators
MWC Mobile World Congress
NAT Network Address Translation
NF Network Functions
NFV Network Functions Virtualization
NFVI Network Functions Virtualization Infrastructure
NIC Network Interface Card
NOMA Non-orthogonal Multiple Access
NV Network Virtualization
O&M Operations & Maintenance
OBSAI Open Base Station Architecture Initiative
OFDMA Orthogonal Frequency Division Multiple Access
OpEx Operational Expenditure
OSPF Open Shortest Path First
OSS Operational Support System
OTN Optical Transmission Network
P2P Peer-to-Peer
PaaS Platform-as-a-Service
PDCP Packet Data Convergence Protocol
PDH Plesiochronous Digital Hierarchy
PE Provider Edge
PGW Packet Data Networks Gateway
PHY Physical Layer
PON Passive Optical Network
POP Point of Presence
PSTN Public Switched Telephone Network
QoE Quality of Experience
QoS Quality of Service
RAN Radio Access Networks
RANaaS RAN-as-a-Service
RAT Radio Access Technologies
RFIC Radio Frequency Integrated Circuit
RLC Radio Link Control
RNC Radio Network Controller
RRC Radio Resource Control
RRH Remote Radio Head
SaaS Software-as-a-Service
SBT Session Border Controllers
SC-FDMA Single Carrier Frequency Division Multiple Access
SDH Synchronous Digital Hierarchy
SDN Software-Defined Networking
SGSN Serving GPRS Support Node
xiv Abbreviations and Acronyms

SGW Serving Gateway

SINR Signal-to-Interference-plus-Noise Ratio
SLA Service-Level Agreement
SON Self-Organizing Networking
SONET Synchronous Optical Networking
SP Service Provider
SSL Secure Sockets Layer
TCO Total Cost of Ownership
TDMA Time Division Multiple Access
TMA Tower Mounted Amplifiers
TP Transmission Point
TTI Transmission Time Interval
UE User Equipment
UTRAN Universal Terrestrial Radio Access
vCPU virtual CPU
vEPC virtualized EPC
vIMS virtualized IMS
VIM Virtualized Infrastructure Management
VLAN Virtual Area Network
VLR Visitor Location Register
vNF virtual Network Functions
vNIC virtualized NIC
VoIP Voice over IP
VPN Virtual Private Network
vRAM virtual RAM
VSN Virtual Sharing Network
WAN Wide Area Network
WiFi Wireless Fidelity
WiMAX Worldwide Interoperability for Microwave Access
WNC Wireless Network Cloud
XaaS X-as-a-Service
List of Figures

Figure 1.1 Global quarterly data and voice traffic in mobile networks
from 2011 and 2016 [1] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
Figure 1.2 Breakdown of the total cost of ownership
(CAPEX and OPEX) for cellular networks [2] . . . . . . . . . . . . . 4
Figure 2.1 Server virtualization: the physical hardware resources
are mapped to multiple virtual machines, each
with its own CPU, memory, disks, and I/O devices . . . . . . . . . 12
Figure 2.2 Network virtualization versus server virtualization . . . . . . . . . . 14
Figure 2.3 Typical VPN topology: a private network deployed
using a public network (usually the Internet) to securely
connect remote sites/users together . . . . . . . . . . . . . . . . . . . . . . 16
Figure 2.4 The ETSI vision of network functions virtualization [3] . . . . . . 21
Figure 2.5 Network functions virtualization: an assessment
of the benefits [4] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
Figure 2.6 Dedicated hosting versus cloud computing
(purchasing IaaS, PaaS, and SaaS) . . . . . . . . . . . . . . . . . . . . . . 27
Figure 2.7 Different service models or layers in the cloud stack . . . . . . . . 28
Figure 3.1 Four principles of SDN. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
Figure 3.2 Mobile backhaul networks (access/aggregation) . . . . . . . . . . . . 38
Figure 3.3 Inter-data enter WAN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
Figure 3.4 Example of controller and switch connection . . . . . . . . . . . . . . 44
Figure 4.1 Virtualized network functions landscape . . . . . . . . . . . . . . . . . . 48
Figure 4.2 ETSI NFV reference architecture . . . . . . . . . . . . . . . . . . . . . . . 49
Figure 4.3 OPNFV architecture framework . . . . . . . . . . . . . . . . . . . . . . . . 50
Figure 4.4 SDN service function chaining architecture . . . . . . . . . . . . . . . 53
Figure 4.5 Optical service function chaining framework . . . . . . . . . . . . . . 54
Figure 5.1 Virtualized PC and SDN routing . . . . . . . . . . . . . . . . . . . . . . . 61
Figure 5.2 Virtualized CPE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
Figure 6.1 LTE request flow latencies [5] . . . . . . . . . . . . . . . . . . . . . . . . . 69
Figure 6.2 Old-style macro BS versus distributed BS . . . . . . . . . . . . . . . . 70
Figure 6.3 Mobile backhaul in 2G/3G/4G networks [6] . . . . . . . . . . . . . . . 73

xv
xvi List of Figures

Figure 6.4 Mobile backhaul for HetNets [7] . . . . . . . . . . . . . . . . . . . . . . . 73

Figure 6.5 The evolutionary role of small cells in cellular networks.
The next and/or current wave of operators’ opportunities
is shown in green . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
Figure 6.6 The 3GPP approaches for eUTRAN sharing . . . . . . . . . . . . . . . 85
Figure 7.1 Future cloud radio access networks (C-RAN) . . . . . . . . . . . . . . 88
Figure 7.2 Cloud RAN [8] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
Figure 7.3 Overall RAN protocol architecture: user and control planes . . . 95
Figure 7.4 LTE protocol stack (downlink) [9] . . . . . . . . . . . . . . . . . . . . . . 96
Figure 7.5 Different C-RAN implementation scenarios
[source Alcatel-Lucent] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
Figure 7.6 Fronthaul transmission (commonly CPRI is used
to connect RRH to BBU) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
Figure 7.7 Key players in C-RAN development
[source: Heavy Reading] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
List of Tables

Table 2.1 Layer 2 VPNs: advantages and disadvantages . . . . . . . . . . . . . . 18

Table 2.2 Layer 3 VPNs: advantages and disadvantages . . . . . . . . . . . . . . 18
Table 2.3 Router versus switch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
Table 2.4 Cloud computing: benefits and risks . . . . . . . . . . . . . . . . . . . . . 30
Table 6.1 Mobile backhaul network primary locations . . . . . . . . . . . . . . . 68
Table 6.2 Distributed BS: features and benefits . . . . . . . . . . . . . . . . . . . . . 71
Table 6.3 Ideal and non-ideal backhaul access technologies . . . . . . . . . . . 75
Table 6.4 The 3GPP network sharing specifications . . . . . . . . . . . . . . . . . 85
Table 7.1 HSPA+, LTE, and LTE-Advanced requirements . . . . . . . . . . . . 103
Table 7.2 Fronthaul throughput for a 3 sector LTE-A site . . . . . . . . . . . . . 104
Table 7.3 Fronthaul (backhaul) compression techniques . . . . . . . . . . . . . . 105

xvii
Chapter 1
Introduction

Over the past decades, the telecommunications industry has migrated from legacy
telephony networks to telephony networks based on an IP backbone. IP-based
networks have offered operators the opportunity to access previously untapped net-
works/subscribers to offer innovative products and services, and stimulate a new wave
of revenue generation. The use of smart phones, tablets, and other data consuming
devices, such as machine-to-machine (M2M) modules, has explosively increased
during past years, and is changing our lives in ways we did not envision. Every day
more people watch more video and run more data-hungry applications using such
devices. New applications are being developed on a daily basis, and M2M devices
are integrated into more areas of life and industry.
Mobile communications experienced a major breakthrough when for the first
time total mobile data traffic topped mobile voice traffic at the end of 2009 [10, 11],
resulting in a paradigm shift from low bandwidth services, such as voice and short
message, to broadband data services, such as video and online gaming. Figure 1.1 [1]
shows the total global data and voice traffic in mobile networks during the past 5 years.
While voice traffic is almost flat, data traffic has experienced a stable exponential
growth. As an example, mobile data traffic has increased nearly 60 Also, mobile data
traffic in the first quarter of 2014 has exceeded total mobile data traffic in 2011 [12].
Mobile networks will face even more increase in data traffic in coming years.
According to Cisco visual networking index forecast [13, 14], by the end of 2016
global yearly IP traffic will pass the zettabyte (1012 gigabytes) threshold, and traffic
from wireless and mobile devices will surpass traffic from wired devices. In addition,
it is projected that by 2018:
• Global Internet traffic will be equivalent to 64 times of that in 2005
• Data center virtualization and cloud computing will grow remarkably and nearly
one-fifth (78%) of workloads will be processed by cloud data centers
• Mobile data traffic will increase 11-fold compared with that in 2013, representing
a compound annual growth rate of 61% between 2013 and 2018
• Busy-hour Internet traffic will grow by a factor of 3.4 (21% faster than average
Internet traffic growth).

© Springer International Publishing AG 2017 1

M. Vaezi and Y. Zhang, Cloud Mobile Networks, Wireless Networks,
DOI 10.1007/978-3-319-54496-0_1
2 1 Introduction

Total monthly traffic (uplink + downlink) in exabytes

6000

5000

4000

3000
Voice
2000 Data
1000

0
Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1
2011 2012 2013 2014 2015 2016

Fig. 1.1 Global quarterly data and voice traffic in mobile networks from 2011 and 2016 [1]

This data flood is caused by different factors. First of all, much more devices will
access Internet and broadband services. There will be nearly 21 billion devices—
including fixed and mobile personal devices and M2M connections—connected to
IP networks [13]; this is 2.7 times as high as the world population, up from 1.7
networked devices per capita in 2013. Being integrated into more areas of life and
industry, M2M devices annually will produce 84% more data, between 2013 and
2018. Furthermore, devices, particularly smart phones, will be more powerful and
services will be more diverse and bandwidth-hungry.
To accommodate the explosively increasing data traffic, many operators see 4G
technologies such as LTE as a way. However, they cannot simply upgrade their net-
work from 2G/3G to 4G overnight because their subscribers may not upgrade their
devices for several years. This implies that operators will need to support multiple
technologies, at least over a period of time, so as to support a mix of subscriber tech-
nologies. More importantly, they need to make a huge investment to upgrade network
infrastructure. As a result, operators are not seeing proportional revenue growth with
the data traffic, and most of them facing flat to declining revenues. Besides, admit-
tedly, even the current 4G networks are not able to address the onslaught of users
demand for mobile data, and network capacity expansion is necessary.
Capacity expansion is one of the most significant technological challenges opera-
tors face today. Despite implementation of advanced capacity-improving techniques
such as multi-antenna systems and increasing use of Wi-Fi offload to increase net-
work capacity, operators can not keep up with exploding capacity needs of customers.
There are multiple other techniques, such as adding spectrum and using smaller cells,
that enable operators to expand their network capacity. These solutions are however,
costly, difficult, and/or take time. Moreover, although bringing up considerable per-
formance gains, they are unlikely to be able to carry the exponentially growing
wireless data traffic. Besides, when a large group of smart phone users gather at
a concert or an arena, for example, heavy usage can exhaust local capacity. Under
1 Introduction 3

such circumstances, traditional radio access network (RAN) architecture is facing

more and more challenges. In addition to the unforeseen capacity challenges, require-
ment for a dedicated equipment room with supporting facilities for each base station
increases network deployment difficulty because cell site costs and space often limit
appropriate locations. Furthermore, power consumption and operating costs go up.
The current RAN architecture is not capable of addressing the explosive need of
data demand; it has run its course. This is partly because it does not efficiently utilize
resources, such as spectrum and processing power, as, with dedicated resources,
certain cells may experience congestion while others are underutilized. Note that
resource allocation in each base station is based on busy-hour traffic (the peak level
of traffic), whereas busy-hour traffic can be much higher than average traffic in many
sites. This is expected to become even worse over time, since busy-hour Internet
traffic is growing faster than average Internet traffic [13, 14]. To address these issues
and meet the growing demand, disruptive solutions are required. Possibly, the gap
will have to be filled by a new generation of networks such as cloud-based networks,
which offer the benefits of cloud computing in RAN [15, 16].

1.1 Challenges of Today’s RAN

Conventional RANs are based on industry-specific hardware owned and operated by

mobile operators. Upgrading technology, reducing ownership costs, and decreasing
carbon dioxide emission at cellular sites are today’s technological, economical, and
ecological challenges for operators around the globe. As elaborated in the following
sections, these issues stem from the architecture of legacy RAN which is based on
dedicated resources for each cell site.

1.1.1 Cost of Ownership

Mobile operators are facing pressure on ever-increasing cost of ownership with much
less growth of income, in the era of mobile Internet. Total cost of ownership (TCO) is
an accounting term that attempts to quantify the direct and indirect financial impact
of a product or system on its life cycle. TCO is designed to look at an investment from
a strategic point of view; it can be divided into two main streams: capital expenditure
(CAPEX) and operation expenditure (OPEX). CAPEX is generally associated with
buying fixed assets or extending the life of an asset and OPEX includes the costs
related to the operation, maintenance, and administration of an investment.
Figure 1.2 illustrates the TCO for cellular systems. In a mobile network, CAPEX
comprises RAN infrastructure (including base station, tower, antennas and feeder,
and microwave) and supplementary equipment, such as power and air-conditioning,
backhaul transmission, and core network, whereas OPEX covers maintenance, cost
of power, rent, and spectrum licence, field-services, planning and optimization.
4 1 Introduction

Fig. 1.2 Breakdown of the total cost of ownership (CAPEX and OPEX) for cellular networks [2]

Roughly speaking, based on several independent estimations or actual cost calcu-

lations, CAPEX comprises one-third of TCO and the other two-third belongs to
OPEX, when considering a period of 7 years for operation. This, however, varies
from network to network.1
Varying based on countries, it is known that 70–90% of CAPEX of a mobile
operator is spent on the RAN [8, 18, 19]. This means that most of the CAPEX
is related to building out cell sites and purchasing the equipment for the RAN.
Then, when we break it down, it appears that the cost of wireless and transmission
equipment, on average, makes nearly 40% of total CAPEX (and more than half of
the RAN CAPEX) [8, 19, 20], and other costs including site acquisition, civil works
and construction, and installation, on average, account for about another 40% of the
total CAPEX of a mobile network. The remaining 20% is spent on other parts of the
network, such as core network and backhaul.
Considering the big share of RAN in total CAPEX and the fact that more than
half of the RAN CAPEX is not spent on wireless equipment, it makes much sense
to lower the expenditure on site construction, installation, and deployment. Hence,
the focus of the mobile operators is to reduce the cost of auxiliary installations and
engineering construction, so as to lower the CAPEX of their mobile networks.

1 For
example, the total U.S. LTE infrastructure OPEX was anticipated to be 57.4 billion, while
CAPEX was projected to be only 37.7 billion between 2012 and 2017 [17].
1.1 Challenges of Today’s RAN 5

1.1.2 Capacity Expansion

The unprecedented growth of mobile data will continue to gain momentum in the
coming years. Adding capacity seems like the obvious answer to meet this challenge;
this is not simple, though. Evolution of wireless technologies has provided higher
spectral efficiency, but it has not been enough and it is difficult to meet this demand
by adding spectrum. Increasing the cell deployment density of the network is another
possible solution, and cell splitting is one of the main trends in RAN that enables
network densification.
Vendors have introduced a wide variety of radio access nodes, such as small
cells and distributed antenna systems, and operators are increasingly augmenting
traditional macro expansion with network offloading solutions. Small cells, used for
increasing capacity and coverage and covering high-traffic public areas, are expected
to account for a large proportion of the offloaded traffic [21]. In addition, by increasing
the cell deployment density of the network, average distances between a user and
the nearest base station decrease. Hence, the link quality improves which results in a
larger capacity for the link [22]. While small cells are viewed as an offload technique
in 3G networks, by introduction of heterogeneous network (HetNet), they are an
integral part of 4G networks.
Although small cells require low power and low cost nodes by definition, today,
they are required to support multiple technologies and multiple frequency bands.
By increasing the implementation of small cells, it is expected to have 50 million
base stations by 2015 [23, 24], and some even predict that in 10–15 years, there
may be more base stations than the number of cell phone subscribers [23]. Hence,
considering cell site costs and space requirements, it is clear that adding capacity
through small cells is difficult and expensive. Small cells will be an integral part of
future networks, but they are not cost-effective, environment-friendly solutions. Nor
are they capable of addressing the long-term mobile data capacity requirements.

1.1.3 Energy Consumption and Carbon Emissions

Information and communications technology (ICT) is one of the major components

of the world energy consumption budget. It is estimated that the global ICT now
accounts for about 10% of the world energy consumption [25–28]. The electricity
consumption of communication networks has been growing by a compound annual
growth rate (CAGR) of 10% per year during 2007–2012, two times faster than the
other sectors in ICT and more than threefold greater than the growth of worldwide
electricity consumption in the same time frame [27, 29]. This is mainly because
during the past few years, operators have been increasingly implementing new cell
sites to offer broadband wireless services, as we mentioned previously, and it was
foreseen to have 50 million base stations by 2015 [23].
6 1 Introduction

With the explosive growth of mobile communications in terms of number of

connected devices, and the demand for new services and ubiquitous connectivity,
the energy consumption of wireless access networks is increasing significantly. In
particular, power consumption rises as more base stations are deployed since it is
estimated that base stations contribute to 60–80% of the total energy consumption
[30]. This situation imposes a major challenge for mobile operators since a higher
power consumption is directly translated to a higher operational expenditures. Car-
bon dioxide (CO2 ) emission is another important consequence of increasing energy
consumption, in addition to rising OPEX. Mobile cellular communication is thought
to account for 2–3% of global CO2 emissions [31]. In 2011, Bell Labs estimated that
mobile base stations globally emit about 18 million metric tons of CO2 per year [32].
This brings about significant environmental impact and is against the current social
and political trend and commitments towards a greener communication.2
The above economical and ecological challenges urge mobile operators to support
RAN architecture and/or deployment scenarios that can cope with the traffic and net-
work growth in a more energy-efficient manner. Cloud RAN, together with software-
defined networking and network functions virtualization, is among the emerging
technologies starting to break the traditional cellular infrastructure in general, and
the radio access network, in particular.

1.2 Cloud RAN - What Is the Big Idea?

Cloud radio access network (C-RAN) architecture is currently a hot topic in the
research, industry, and standardization communities. The basic idea behind C-RAN
is to change the traditional RAN architecture in a way that it can take advantage
of technologies like cloud computing, wireless virtualization, and software-defined
networking. More specifically, C-RAN is a RAN architecture in which dedicated cell
site base stations are replaced with one or more remote clusters of centralized virtual
base stations, each of which is able to support a great many remote radio/antenna
units. C-RAN may also stand for centralized RAN. Centralized RAN is basically an
evolution of the current distributed base stations, where the baseband unit (BBU)
and remote radio head (RRH) can be spaced miles apart. Centralized RAN and cloud
RAN can be considered as two sides of the same coin. Although some people may
use these two terms interchangeably, there is a clear difference between them. Cloud
RAN implies that the network is “virtualized” on top of being centralized, meaning
that it is implemented in generic server computers (or blade servers) and base station
resources can be added as per their needs, to efficiently handle the network traffic.
Depending on the function splitting between BBU and RRH, cloud RAN partly or
wholly centralizes the RAN functionality into a shared BBU pool or cloud which is
connected to RRHs in different geographical locations.

2 As an example, the UK is committed to reducing the amount of CO2 it emits in 2050 to 20% of
that seen in 1990 [31].
1.2 Cloud RAN - What Is the Big Idea? 7

1.2.1 Advantages

Cloud RAN has strategic implications on operator–vendor relationship, as it allows

operators to implement network upgrades more agilely and to select between vendors
easily. Aside from the strategic implications, the C-RAN architecture has practical
and measured benefits to the current RAN, that essentially revolved around reducing
the cost of network operations. We briefly review some of them here.
Major Savings in CAPEX and OPEX
From a business perspective, C-RAN is expected to bring in significant reductions
in both CAPEX and OPEX due to reduced upgrading and maintenance costs. Cost
saving in CAPEX is due to the fact that single cells are not required to be dimen-
sioned for peak-hour demands. Instead, baseband processing power can be pooled
and assigned specifically where needed, implying that dimensioning can be done for
a group of cells rather than a single one. This increases the processing utilization
largely. Also, baseband processing can be cost effectively run on commercial servers.
Further CAPEX savings can be achieved from potential technology enhancements
(e.g., LTE-Advanced features) which leave further processing headroom. In addition,
less costly general-purpose processor hardware can be used for RAN algorithms.
OPEX savings can be drawn mainly from energy savings, reduced cost of mainte-
nance, and smaller footprint required. Generally, operations and maintenance (O&M)
for distributed hardware (conventional RAN architecture) is more costly than that
of centralized hardware (cloud RAN architecture). Also, with C-RAN centralized
network analysis and optimization, such as centralized self-organizing networks can
be naturally evolved which are able to transform network economics. Besides, due
to smaller footprint required at each site, site rental and civil works costs drop.
Flexibility in Network Capacity Expansion
From the network capacity expansion point of view, C-RAN brings in significant gain
and flexibility. In a heterogeneous network C-RAN, low power RRH can be deployed
in the coverage area of a macro cell where a high level of coordination between the
macro cell and the RRH is achievable. This can reduce interference when some LTE-
advanced technologies such as coordinated multipoint (CoMP), where multiple base
stations transmit and receive from a mobile device, are deployed.
Reducing Energy Consumption and CO2 Emissions
From an ecological perspective, C-RAN architecture is preferred to the conven-
tional RAN architecture, as it consumes less energy and is greener. This is mainly
because multiple BBUs can share facilities, e.g., air-conditioning, and partly because
of resource aggregation which results in an improved resource utilization, which in
turn improves energy efficiency. On the same page, in the centralized architecture, if
required, e.g., when traffic demand is low, BBU resources can be switched off much
more easily than the conventional distributed RAN. This brings in further energy
savings. It is obvious that such an architecture is more environmental friendly and
reduces CO2 emissions.
8 1 Introduction

In addition to the above-mentioned benefits, adopting C-RAN allows different lev-

els of sharing in access network. It allows the operator to efficiently support multiple
radio access technologies (RAT), network sharing (sharing base band processing,
RRH, and spectrum), or outsourcing. It should be mentioned that the implementa-
tion of a centralized RAN is easier than a cloud RAN. It, however, lacks the benefits
associated with virtualization and cloudification.

1.2.2 Challenges

Although the basic ideas of C-RAN are already relatively mature, we are still in its
early days and much work is yet required to achieve this vision. There are many
challenges such as fronthaul (between BBUs or RRHs) requirements in terms of
bit rate, latency, jitter and synchronization, interface definitions between BBUs and
RRHs and between BBUs, and base station virtualization technology. Besides, cur-
rent general-purpose processors (GPPs) are not a practical solution for handling the
datapath processing and the very high data rates required by 4G systems. Also, these
GPPs are not optimal platforms for such operations in terms of power consumption.

1.3 Related Technologies

Virtualization is a key enabler of cloud computing and cloud-based infrastructures.

There are also other emerging technologies, such as network functions virtualization
(NFV) and software defined networking (SDN), that support cloud environments.
These technologies move the networking industry from today’s manual configuration
to more automated and scalable solutions. They are complementary approaches that
solve different subsets of network mobility problem.
Wireless carriers are targeting the integration of NFV and SDN across mul-
tiple areas including radio access network, core network, backhaul, and opera-
tional/business support systems (OSS/BSS), caused by the promise of total cost
of ownership reduction. SDN and NVF are among other initiatives to move from the
traditional cellular infrastructure toward a cloud-based infrastructure where RAN,
mobile core, etc. are expected to be applications that can run on general-purpose
infrastructure, rather than proprietary hardware, hosted by data center operators and
other third parties. We briefly explain these enabling/related technologies and their
relation to C-RAN in the following sections.
1.3 Related Technologies 9

1.3.1 Network Virtualization

Virtualization is a technology that enables us to go beyond the physical limitations

normally associated with entire classes of hardware, from servers and storage to net-
works and network functions. Network virtualization (NV) ensures that network can
integrate with and support the demands of virtualized architectures. It can create a vir-
tual network that is completely separate from other network resources. Virtualization
mechanisms are at the core of cloud technologies.

1.3.2 Network Functions Virtualization

Network functions virtualization (NFV) provides a new way to design, deploy, and
manage network services. It decouples the network functions from purpose-built
hardware, so they can run in software. Therefore, NFV enables the implementation
of services on the general-purpose hardware, allowing operators to push new services
to the network edge, i.e., to base stations. This in turn helps operators support more
innovative location-based applications and reduces the backhaul traffic by shifting
services away from the network core.

1.3.3 Software-Defined Networking

Software-defined networking (SDN) is a new approach to designing, building, and

managing networks which enables the separation of the network’s control plane and
data plane, which makes it easier to optimize each plane. SDN has the potential
to make significant improvements to service request response times, security, and
reliability. In addition, by automating many processes that are currently done man-
ually it could reduce costs. SDN is a natural platform for network virtualization
as, in a software-defined network, network control is centralized rather than being
distributed in each node [33]. Therefore, this technology can be applied to C-RAN
environments to enable universal management capabilities, allowing operators to
remotely manage their network.
In summary, NV, NFV, and SDN each provide a software-based approach to net-
working, in order to make networks more scalable and innovative. Hence, expectedly,
some common beliefs guide the development of each. For example, they each aim to
move functionality to software, use general-purpose hardware in lieu of purpose-built
hardware, and support more efficient network services. Nevertheless, note that SDN,
NV, and NFV are independent, though mutually beneficial. Finally, by applying the
concepts of SDN and NFV in a C-RAN environment, most of the processing can be
implemented in commodity servers rather than proprietary appliances. Hence, when
combined with SDN and NFV, C-RAN provides operators with reduced equipment
costs and power consumption.
 10 1 Introduction

1.4 Outline of Chapters

This book is divided into five chapters and provides information on the different
technologies enameling cloud RAN.
Chapter 1: Introduction. In this chapter, we have introduced the challenges of
conventional radio access network (RAN) and the requirements for future RAN. This
was followed by a brief overview of cloud RAN and its advantages. We have also
outlined the developing technology relevant to cloud RAN.
Chapter 2: Wireless Virtualization. This chapter reviews various wired network
virtualization technologies as well as network functions virtualization. It then studies
the state of the art in the wireless virtualization and its advantages and challenges.
The last part of this chapter is devoted to the cloud computing, its service models,
and its relation to virtualization.
Chapter 3: Software Defined Networking. In this chapter, we provide a review
of the SDN technology and business drivers, describe the high-level SDN architecture
and principles, and give three scenarios of its use cases in mobile access aggregation
networks and the cloud networks. Furthermore, we provide discussions on the design
implementation considerations of SDN in the mobile networks and the cloud, in
comparison with traditional networks.
Chapter 4: Virtualizing the Network Services: NFV. In this chapter, we provide
a survey of the existing Network Function Virtualization (NFV) technologies. We
first present its motivation, use cases, and architecture. We then focus on its key use
case, the service function chaining, and the techniques and algorithms.
Chapter 5: SDN/NFV Telco Case Studies. In this chapter, we review the two
important case studies of SDN and NFV’s usage in telecom mobile network. In
particular, we discuss network virtualization’s usage in packet core network and
in customer premise equipment (mobile edge networks). In both case studies, we
discuss the challenges and the opportunities.
Chapter 6: RAN Evolution. The main objective of this chapter to set the stage
to better understand and needs for the cloud RAN architecture, in Chap. 7, and its
barriers and/or competing technologies. The chapter starts with an overview of the
architecture of mobile networks with an emphasis on RAN and backhaul/fronthaul
solutions. It then compares the legacy and distributed base stations technologies. It
also describes the current and emerging trends in wireless network densification as
well as several concepts related to the cloud RAN solution, including small cell,
distributed antenna systems, and mobile network sharing.
Chapter 7: Cloud RAN. In this chapter, first the cloud RAN is defined, its driving
concepts are elaborated, and its vision and mission are identified. Then, different
implementation scenarios for the cloud RAN are studied in detail and compared
with. Next, the conclusions are drawn on the possible solution for future networks
with a view on the competing solutions such as small cells and edge cloud. Finally,
the challenges and research frontiers are identified and described.

View publication stats