5Connectivity Resources Management in [601328]

5Connectivity Resources Management in
5G networks

5G systems

A digital transformation, brought about through the power of connectivity, is taking place in almost
every industry. Through an unprecedented ability to share information, people and industries are
collaborating more, creating solutions that combine many different areas of expertise and overturning
traditional business models. But the pressure on the networks that provide this con nectivity is palpable.
For that reason, 5G systems will be built to enable logical network slices, which will enable operators to
provide networks on an as -a-service basis and meet the wide range of use cases that the 2020
timeframe will demand.
And so, tr aditional cellular networks and their one -size-fits-all approach need to be adapted so that the
thousands of use cases, many different subscriber types and varying app usage can be supported. [1]
Evolving existing radio -access technologies (RATs), like LTE , and new 5G technologies will all be part of a
future flexible and dynamic 5G system. Support will exist for cross -domain integration and multi -RAT
environments. [1]New technologies will enable concepts like very low latency, and the need for
additional c apacity will require communications to operate in higher frequency ranges than they do
today.
Services

The expectations for 5G networks are high – providing support for a massive range of services. Industry
transformation, digitalization, the global depe ndence on mobile broadband and the rise of innovative
industrial applications all require new services, which has a considerable impact on the transport
network .

5G radio

How the 5G radio is deployed determines the level of flexibility needed in the transport network.
Capacity, multi -site and multi -access connectivity, reliability, interference, inter -site coordination, and
bandwidth requirements in the radio environment place tough demands on transport networks.
In 5G, traditional macro networks mig ht be densified, and complemented through the addition of small
cells. Higher capacity in the radio will be provided through advances in radio technology, like multi -user
mimo and beamforming, as well as the availability of new and wider spectrum bands [2 ]. Consequently,

the capacity of the 5G radio environment will reach very high levels, requiring transport networks to
adapt. Not only will transport serve a large number of radio sites, but each site will support massive
traffic volumes .

For example, a UE that is connected to a number of sites simultaneously, may also be connected to
several different access technologies. The device may be connected to a macro over LTE, and to a
small cell using a new 5G radio -access technology. Multi -site and multi -rat c onnectivity provides
greate r flexibility in terms of how UE s connect to the network and how e2e services are set up across
radio and transport. For example, allowing for efficient load balancing of UEs among base stations not
only improves user experience, it also improves connection performance.

The impact of interference may favor deployment models where coordination can be handled more
effectively. In small -cell deployments, UEs are often within reach of a number of base stations, which
increases the l evel of interference, and at times requires radio coordination capabilities for mitigation.
However, the method used for handling interference depends on how transport connectivity is
deployed. In a centralized baseband deployment, tight coordination featu res, such as joint processing,
can be implemented. In traditional Ethernet and IP-based backhaul, tight coordination requires low –
latency lateral connections between participating base stations.

Centralized baseband processing tends to result in lower op erational costs, which makes this approach
interesting. However, it typically comes at the cost of high bandwidths in the transport network. The
high bandwidth, together with stringent delay and jitter requirements, makes dedicated optical
connectivity a p referred solution for front haul.

In 5G networks, the bandwidth requirements for front haul could be very high. The demand will be
created by, for example, antennas for mu -MIMO and beam forming – which could use in the order of
100 antenna elements at each location. But some primary networking principles remain valid, such as
timing and synchronization. Defining new packet -based front haul and mid haul interfaces requires the
underlying network to include protocols and functions for time -sensitive trans port services. Related
standardization e fforts are currently underway [3 ].
SPECTRUM IN THE 5G ERA

5G systems will need to provide significant improvement in cell capacity to accommodate the rapidly
increasing traffic demands. Although 5G will introduce a bevy o f new technologies that enable networks
and devices to make better use of scarce spectrum resources, more efficient use of current spectrum
resources will not be sufficient to keep pace with the mobile data usage increase. 5G systems are
expected to provid e data rates on the order of gigabits per second, anytime and anywhere. This could
only be realized with much more spectrum than that currently available to IMT2 systems through the
International Telecommunication Union’s (ITU) process. Frequency bands cur rently in use by IMT
systems are fragmented with varying degrees of availability and amount of bandwidth across bands,
countries, and regions, leading to problems such as roaming, device complexity, lack of economies of
scale, and harmful interference. Som e technologies for aggregation of IMT bands have been developed,
but they have limitations in meeting the broader bandwidth needs of future systems. Therefore,
contiguous and broader frequency bands are needed to provide gigabit data rate services in the f uture.
All spectrum currently available to cellular mobile systems, including IMT, is concentrated in bands
below 6 GHz due to the favorable propagation conditions in such bands.

For the same reason, these frequencies are in high demand by other services, including fixed,
broadcasting, and satellite communications. As a result, these bands have become extremely crowded,
and prospects for large chunks of new spectrum for IMT below 6 GHz are not favorable. Recent
advancements in mobile communication systems and devices operating at frequencies around 60 GHz,
combined with advancements in antenna and RF component technologies, have opened the gates to
using non -conventional bands for cellular applications. Such advancements will help enable dense small
cell deployments over a diverse set of spectrum that includes higher unconventional bands. Such
deployments will be an important 5G usage scenario as there will be continued need to meet
exponential growth in traffic demand and address the requirement for gigabit data rates everywhere,
including at the cell edge. It is expected that small cells operating over spectrum not traditionally used
by cellular systems (e.g., mm -wave bands, such a s bands around 30 –50 GHz) will be deployed indoors or
outdoors to meet this requirement toward “edgeless” 5G networks. Another potential means to make
large amounts of spectrum available to future IMT systems is through novel schemes such as LSA,
whereby c ertain underutilized non -IMT spectrum could be integrated with other IMT spectrum in a
licensed pre -determined manner following mutual agreement among the licensees. In this way, LSA has
the potential to expand some existing IMT bands and make them suitabl e to support channel bandwidth
larger than is possible today. As the mobile industry continues its efforts to open new technological
frontiers, governments worldwide can also play a role in meeting growing data demand by adopting
sensible, innovative, and technology -neutral spectrum policies to facilitate more efficient use of
spectrum, and thus create new economic opportunities.

Network architecture (C-RAN and SDN )

Cellular -based network architecture design, from traditional circuit switch (CS) + packet switch (PS)
2G to all -IP flat 4G, enables mobile communications to achieve unprecedented success. Today, the
cellular networks have evolved into a huge multi -radio access technology (multi -RAT) and multi -layer
heterogeneous network. However, in face of emerging mobile internet applications and digital floods,
this architecture becomes more and more incompetent. Traditional single -RAT base station deployment
becomes an unbearable cost, O&M burden to operators. Therefore, some state -of- art techniques, such
as cloud and software defined network (SDN) [4 ], are being introduced to mobile networks. The
architecture that is operating on cloud RAN (C -RAN) [ 5-7] and SDN [ 7-9] attracts great attention from
both academic and industrial partners .

According to the above resear ches, new network architecture is presented in Figure 1. The architecture
consists of cloud application , SDN controller cloud, SDN -based C -RAN, SDN -based transport network,
and SDN -based core network [ 9, 10]. Application cloud provides various services, su ch as network
management and performance monitor etc. SDN controller cloud transforms policies from application
cloud and provides centralized control services for objective e -network elements by these policies, such
as C-RAN, transport network, and core network . The SDN -based C -RAN consists of large -scale baseband
unit (BBU) pools, r adio -over -fiber (RoF) systems [4 ], and distributed radio access points (RAPs) and light
RAPs (LRAPs).

The BBU pools give brought together baseband signal handling a long ways past single base station.
RAPs accomplish flagging scope like large scale cell, while LRAPs accomplish information transmission
like little cell The RAPs and LRAPs associate BBU by me ans of the RoF framework. The RoF framework
accomplishes astute associations between RAPs/LRAPs and BBU. The SDN controller gives dynamical
data transfer capacity modification of every RAP/LRAP to BBU association and BBU pools administration
including syst em RAT. The SDN -based transport system accomplishes adaptable backhaul systems
administration and determination, dynamical transport transmission capacity alteration of each RAN to
center system association. The SDN -based center system comprises of unifie d control entity (UCE) and
unified data gateway (UDW).

UCE achieves unified control function, which integrates mobility management entity (MME), service
gateway control plane (SGW -C), and packet data network gateway control plane (PGW -C). UCE together
with SDN controller manages general packet radio service (GPRS) tunneling protocol for user plane (GTP –
U) tunnels. UDW achieves data forwarding function, which integrates service gateway data plane (SGW –
D) and packet data network gateway data plane (PGW -D).

The cloud system has a few advantages as takes after:

(1) The expansive incorporated sending pro -videos processing capacity better than single arrangement;

(2) Virtualization empowers the sensible partition of programming and equipment, which diminishe s
unpredictability of handling and extends limit of equipment;

(3) New administration answers for dense system deployment , for example, radio access nework as an
administration (RANaaS) [9]. In the design, application server, SDN controller, RAN and cor e system
embrace the cloud arrangement. The SDN decouples core control from individual e quipments and

relocates it into accessible computing devices, which empowers the hidden base to be preoccupied for
applications and system administrations, i.e., the sy stem is dealt with as a legitimate or virtual entity[6].
Along these lines, the SDN can rapidly react to changing system nodes, business needs or client requests
and so on.

The SDN architecture consists of application layer, control layer, and infrastructure layer. The interface
between application layer and control layer is open application programming interfaces (APIs), while it is
control/data plane interface between control layer and infrastructure layer. In the architecture, the
control and data plane of RAN and core network are separated. For RAN, RAPs achieve signaling
coverage to maintain user connection like macro cell, while LRAPs achieve data transmission like small
cell. RAPs equipped massive MIMO can also provide data transmission. B ecause L RAPs are closer to
users, it im proves channel condition to achieve better performance .[8]

The SDN engineering comprises of application layer, control layer, and infrastructure (base) layer . The
interface between application layer and control layer is open application programming interfaces (APIs),
while it is control/information plane interface between control layer and base layer. In the architecture ,
the control and information plane of RAN and center system s (core) are isolated. For RAN, RAPs
accom plish signaling scope to keep up client association to macro cell, while LRAPs accomplish
information transmission like small cell.[6,8] RAPs are prepared with gigantic MIMO that can provide
information transmission. Since LRAPs are closer to users , it enhances channel condition to accomplish
better execution.
CONCEPTS AND TECHNOLOGIES FOR RAN 5G NETWORK ARCHITECTURE

The most promising enabling technologies that are expected to be adopted in 5G network architectu re,
namely C -RAN, Network Functions Virtualization (NFV) and Software -Defined Networking (SDN), to cater
the requirements identified above, especially to reduce the operation cost from the perspective of
operators. Beyond that, the relations among C -RAN, N FV and SDN are summarized, leading directly how
C-RAN integrates the other two technologies to realize multiple levels centralized management in BBU
pool.

C-RAN , i.e. Cloud -RAN, also referred as Centralized RAN, was first introduced by China Mobile Rese arch
Institute in April 2010 [ 11]. Simply speaking, C -RAN is a centralized, cloud computing based new cellular
network architecture that has the ability to support current and future wireless communication
standards. By separating the Baseband Units (BBUs) from the radio access units, and migrating BBUs to
the cloud forming a BBU pool for centralized processing, C -RAN provides several advantages compared
to the conventional RANs, such as scalability and flexibility of further deployment of plenty of remote
radio heads (RRHs).

NFV refers to the implementation of network functions in software running on general purpos e
computing/storage platforms [12 ]. This approach allows the deployment of network functions in data
centers to leverage the traffic load throug h virtualization techniques. Compared to the conventional
network which implements network functions on dedicated and application specific hardware, the main
motivation for NFV is to reduce life and innovation cycles within telecommunication networks throu gh
software update rather than hardware update, thereby leveraging from the economy of high volume
hardware platforms. NFV has recently attracted significant interests from the industry, which has led to
the creation of a dedicate industry study group at E TSI [ 13].

SDN is an emerging architecture that decouples the network control and forwarding functions, enables
the network control to become directly programmable and the underlying infrastructure to be
abstracted to adjust applications and network servi ces dynamically [ 4]. Based on SDN architecture, the
network intelligence is (logically) centralized in software -based SDN controller that maintains a global
view of the network, which appears to applications and policy engines as a single and logical swit ch, and
enables network managers to configure, manage, secure, and optimize network resources very quickly
via dynamic, automated SDN programs, by which they can rewrite themselves due to the independence
of the programs on the proprietary software. SDN si mplifies network design and operations because
instructions are provided by SDN controllers instead of multiple, vendor -specific devices and protocols.
These three concepts relate to each other, but contribute for different domain. SDN is focused on the
separation of the network control layer from its forwarding layer, while NFV is focused on implementing
network functions in software on standard IT platforms in order to enable the migration from
proprietary appliance based embodiments to a standard hardwar e and cloud based infrastructure. In
mobile networks, SDN and NFV are currently discussed in the context of vitalizing the core network and
rarely involved in RAN [12]. First trials are performed demonstrating that critical mobile network
functions such as MME, HGW/PGW, or HSS can be implemented on standard IT platforms based on SDN
and NFV technologies [ 11,14].

In case of RANs, when NFV is used for logically centralizing the base band processing within the RAN, C –
RAN becomes an application of NFV. Three for ms of virtualization in a C -RAN solution can be potentially
co-existed on a cloud BBU pool, based on General Processing Processor (GPP) platform: hardware
virtualization, network virtualization and application virtualization [14]. – In hardware virtualizat ion, a
portion of the RAN hardware is virtualized, i.e., a hypervisor is used to create and manage a virtual RAN
layer. The hypervisor may run on/or on the top of the Real Time Operating Systems (RTOSs). The layer
that is virtualized maybe a portion of the physical (PHY) layer or a protocol oriented layer suc h as the
MAC layer and above. In network virtualization, network elements such as switches, routers, edge
caching storage elements and transport resources are abstracted and combined into a pool that is
managed by a network operating system. These virtual network components can then be assigned and
managed by a u nified network control policy. [14,15 ]

Application virtualization is also within the RAN, where the network management application entities
are repl aced by a virtualization layer which allows existing applications to run in the C -RAN without the
need to rewrite the software. Following the discussions above, on one hand, C -RAN can be seen as an
application of NFV and SDN, on the other hand, C -RAN enabl es both full centralization and distribution
of RAN functions, this needs not to be the case with a general NFV implementation where only a subset
of all modules may be implemented centrally or the radio access points implement all functions based
on gener al purpose hardware [15]. Therefore, the C -RAN architecture integrated both SDN and NFV
technologies has the scalability and flexibility to adjust the development of future mobile networks, and
its cost effective solutions to face the challenging scenarios of 5G need further research.

C-RAN OVERVIEW

This section describes deployment architectures and potential so lutions of the C -RAN systems. The
general architecture of C -RAN is illustrated in Fig 2, which consists of three main comp onents, namely
BBU poo l with centralized processors, RRHs with antennas located at the remote sites, fronthaul
network which connects the RRHs to the BBU pool and requires high bandwidth and low -latency.

FIG. 2 C-RAN ARCHITECTURE [15]

The RRHs transmit the RF signals to UEs in the downlink or forward the baseband signals from
UEs to the BBU pool for further processing in the uplink. In general, the RRHs include RF
amplification, up/down conversion, filtering, A/D and D/A conversion and interface adaptation.
By conducting most of signal processing functions in the BBU pool, RRHs can be relatively
simple, which can be distributed in the large network in a cost -efficient manner.

The BBU pool is composed of BBUs which operate as virtual base stations to process baseband
signals and optimize the network resource allocation. Considering different demands on
network performance and system implement complexity, the BBU assignment for each RRH
coul d be implemented in a centralized or distributed manner depending on different resource
mana gement in BBU pool distributed manner, one RRH directly connects to its exclusive BBU.
This manner is simple and easy to be realized. [15,16 ]But it is not beneficial to exploit the
advantages of joint signal processing and central controlling in C -RAN. In the centralized
manner, all RRHs may connect to a switcher/central device, which can flexibly schedule
processing resource in BBU pool for one RRH or a set of RRHs. This manner features many
advantages in term of flexibly resource sharing and efficient energy efficiency by joint
scheduling. Beside, centralized processing permits the implementation of efficient interference

avoidance and cancelation a lgorithms across multiple cells. It provides the means to selectively
turn RRHs on/off in line with the traffic fluctuations in different scenarios.

The fronthaul links can be realized by different technologies, like fiber or wireless way, and falls
into two categories: ideal without bandwidth constrain and non -ideal. An overview of possible
wired and wireless backhaul technologies for the C -RAN architecture is given in [16] and [17].
Optical fiber is always seen as a way of ideal fronthaul for C -RAN, because it can provide large
bandwidth and high data rate. For instance, the NG -PON2 has the following performance
specifications: bandwidth of 40 GHz and 10 Gbps for the do wnstream and upstream
respectively, and range up to 40 km. VDSL2 utilizes up to 30 MHz of bandwidth to provide
speeds of 100 Mbps both downstream and upstream within 300 -400m. Wireless backhauls are
faster and cheaper to deploy than fiber. Traditional wire less backhaul generally operates on the
licensed band by reuse techniques such as relay. Wireless backhaul can also employ the
microwave technology with carrier frequencies between 5 and 40GHz. However, due to the
limited available bandwidth at these frequ encies, they can provide data rates of only a few
hundred Mbps . [18]

Allocation RRH -BBU Functions . Because the ideal fronthaul link is not available everywhere,
some function split options between RRH and BBU are presented, which will have a big impact
on the details of C -RAN architectures [14]. Although there are various possibilities, three cases
are listed as shown in Fig. 2 Option 1: the L1 (Layer 1, PHY), L2 (Layer 2, MAC) and L3 (Layer 3,
network) functions all implemented in BBU. This option is the premi er C-RAN configuration,
which is clear and simple. The disadvantage is high burden on fronthaul, due to a large amount
of data from RRH. Option 2 is moving the L1 function to RRH, which greatly reduce the RRH -BBU
data load since L1 takes major computation parts of RAN. However, some advanced features
such as Coordinated Multiple Point Transmission and Reception (CoMP), joint processing
Distribute Antenna System (DAS), cannot be efficiently supported. The interaction between
MAC and PHY could also be complex . And it increases the difficulty of interconnection between
L1 and L2. Option 3 is stripping partial L1 function (like user specific or cell specific) from each
BBU, and assembling them into one processing unit, which may be a part of the BBU pool. The
benefit of option 3 is its flexibility to support resource sharing and the potential to reduce
energy consumption in BBU. There exist other alternatives to allocate RRH -BBU functions
considering the tradeoff between the burden on fronthaul and the flexibilit y of resource
schedule [ 8], [17].

C-RAN IMPLEMENTATION IN 5G MOBILE NETWORKS

UDN and coexistence of Multi -RATs are significant scenarios of 5G mobile networks. In this section we
will discuss the challenges and potentials for C -RAN implementation in aforesaid scenarios. A.
CHALLENGES OF C -RAN UNDER UDN DEPLOYMENT Small cell deployment is one of the most promising
technological trends to cope with the rising need for very high data rates foreseen in future mobile
networks which involve various types o f small cells such as pico cell, relay, Machine -to-Machine, and
also Device -to-Device [ 19].

Fig. 3. C-RAN ar chitecture under UDN deployment [ 15]

As Fig. 3 illustrat es, because of the densification of small cells, the distance between UE and access
point is greatly reduced, i.e. tens of meters, therefore one UE will fall into several small cell coverage ,
and all of those small cell BSs can receive the UL signal from th e UE or serve the UE in downlink. Based
on this property of UDN, some joint/collaboration mechanisms can provide better experience of users.
In traditional cellular networks, the cooperation among cells needs a large amount of signaling
exchange, and alway s do not perform as well as expected under practical constraints due to backhaul
delay, backhaul capacity and user mobility. By the aid of C -RAN, the information exchange and central
decision are able to be performed in BBU pool, which brings performance g ain from cooperative
processing. Besides, the non -ideal characteristics of fronthaul link require a joint design with BBU
processing. Previous work in [19] considers using adaptable techniques including multipoint Turbo
detection, coordinated beamforming a nd joint network -channel coding in PHY layer. In MAC layer,
considering real -time schedule and fronthaul constrains of delay and throughput, smaller time -scales
algorithms are proposed. Based on these approaches, a joint design of BBU processing and fronth aul
network can be effectively carried out, in order to enable the deployment of small -cells and
heterogeneous fronthaul [ 20].

Frequent handover will occur under UDN, which will bring high burden to the signaling overhead and
harm user experience. Beside s, handover failure and ping -pong issues may also appear by reduced small
cell size and larger intercell interference. With BBU centralization in C -RAN, handover can be realized in
a simplified procedure and improved performance can be achieved. In [39] th e handover performance
of C-RAN is analyzed over a baseline decentralized RAN for Global System for Mobile Communications
(GSM), Universal Mobile Telecommunication System (UMTS) and Long Term Evolution (LTE) systems.
The results indicate that, lower total average handover interrupt time can be achieved in GSM thanks to
the synchrono us nature of handovers in C -RAN [15,20 ]. For UMTS, inter -NodeB soft handover would
become intra -pool softer handover in C -RAN. This brings some gains in terms of reduced signaling, less
Iub transport bearer setup and reduced transport bandwidth requirement.

For LTE X2 -based inter -eNB handover, C -RAN could reduce the handover delay and to a large extent
eliminate the risk of UE losing its connection with the serving cell while still wait ing for the handover
command, which in turn decreases the handover failure rate. Under UDN deployment, there is a
discernible phenomenon that traffic load varies from time to time such as the typical night -day behavior
of users and their daily swarming to offices and back to residential areas. While traffic varies, the power
consumption of the radio access network does not effectively scale with it, so small cells switch on/off
schemes are proposed to dynamically accommodate the traffic which causes changes in the network
topology in an energy effi cient manner.

In C-RAN architecture, the BBU pool can evaluate the current situation of whole network in terms of
traffic demand, system performance level, target system performance level, criticality of user, en ergy
status at the time of day and many other similar factors to decide whether the activation of the
respective small cell RRH is needed or not. For example, if a certain part of the macro cell contains a
large number of UEs and the rest of the cell area has a low number of users with low traffic demand
requests, the BBU pool can offload the users in the congested area by activating the respective RRHs
and handle the rest of the users by macro RRH. [15] This will incur less handovers and thus consume less
energ y compared to the case that the total area is always covered by small cells. On the other side, it
will bring higher capacity compared to the case that the total area is covered only by macro cells. This
dynamic approach enables the network topology to cha nge based on current demand levels and
performance expectations.

POTENTIAL ISSUES FOR C -RAN

FRONTHAUL ISSUES

As one of the promising evolution path for future cellu lar network architectur e, C-RAN has attracted
many aca demic research interests. Through the ability of centralized
processing/controlling/decision in C -RAN, the joint pro cessing becomes relatively easy
joint scheduling, joint interference alignment/cancellation and other more advanced future
technologies, such as multi -RAT support, common O&M reliability and network sharing,
which are diffi cult to be implemented in curr ent cellular networks [ 12], [15].

However, some joint processing techniques envisioned for such cloud -based networks require RRH to
deliver received signals to BBU pool, usually in the form of in -phase and quadrature -phase components
(I/Q-stream), which demand a drastically higher data rate on the fronthaul link than decoded
user data [ 15]. Until recently, the required data rate could only be provided by ber and not by wireless
links, which could be reached more easily and at lower cost in C -RAN architecture [ 17]. There are two
kinds of efforts are beind done to solve the problem: one is to reducing the through put requirement on
transportation links by means of joint coding under a data rate constraint , or resampling, rescaling
and removing some u nnecessary parts of the signal , [21 ] the other is developing enhanced fronthaul
technology that supports wideband and high data rate, such as millimeter radio links [17].

TRAFFIC STEERING AND SERVICE MAPPING

Additionally, the notion of intelligence is being exploited so as to guarantee the ubiquitous and reliable
user experience and improve performance of Multi -RAT integrated deploy ments by taking advantage of
global information about users, services and network status.. The inputs to this intelli gence process
comprise information on the traffi c, mobility levels, and interference levels regarding transceivers and
physical elements on the cloud. Intelligence shall then lead to decisions related to:

1) the cell confi guration and resource allocation, and the trafi c distribution to the cells;

2) the assignment of functional components to the computing resources;

3) the best connections between RRH and BBU pool and the best interconnections betwe en the multi –
RAT operations.

Various users have diverse mobility and service properties, realized by different process inside the
wireless system, such as Intelligent Copying S ystem without mobility and realtime traffi c demand,
whose traffi c can be assigne d in night when wireless traffi c load is low. Besides, no mobility related
policies are expected to be launched for this class user. While for smart terminal user, pro mpt service
access and QoS guar antee are both needed. Therefore, the traffi c steering policies considering users and
service properties are the fir rst step of intelligence on traffi c steering and service mapping function.

REFERENCES

[1] – 5G systems, Eri csson White paper Uen 284 23 -3244, January 2015
[2]- Ericsson Review, June 2014, 5G radio access, available at: http://www.ericsson.com/news/140618 –
5g-radioaccess_ 244099437_c
[3] – IEEE, 802.1CM – Time -Sensitive Networking for Fronthaul
[4] – ONF. White Paper on Software -Defined Networking: the New Norm for Networks, 2012
[5] – China Mobile Research Institute. White Paper on C -RAN: the Road Towards Green RAN Version 2.5,
2011
[6] – Chang G K, Liu C, Zhang L. Architecture and applications of a versatile sm all-cell, multi -service cloud
radio access network using radio -over -fiber technologies. In: Proceedings of IEEE International
Conference on Communications Workshops, Budapest, 2013.
[7] – Rost P, Bernardos C J, Domenico A D, et al. Cloud technologies for flexible 5G radio access networks.
IEEE Commun Mag, 2014
[8] – Bernardos C J, de la Oliva A, Serrano P, et al. An architecture for software defined wireless
networking. IEEE Wirel Commun, 2014, 21: 52 –61
[9] – Costa -Requena J. SDN integration in LTE mobile backhaul networks.
[10] – Ben Hadj Said S, Sama M R, Guillouard K, et al. New control plane in 3GPP LTE/EPC architecture
for on -demand connectivity service.
[11] China Mobile Research Institute. (Apr. 2010). The 1st C -RAN International Workshop. [Online] .
Available: http://labs.chinamobile. com/focus/C -RAN
[12] – FP7. (Dec. 2013). Future Networks Cl uster 5G Radio Network Architec ture. [Online]. Available:
http://cordis.europa.eu/fp7
[13] – Network Functions Virtua lization: An Introduction, Bene ts, Enablers, Challenges and Call for
Action. [Online]. Available: http:// portal.etsi.org/nfv/nfv -white -paper.pdf
[14] – A. Basta, W. Kellerer, M. Hoffmann, K. Hoffmann, and E. -D. Schmidt, „A virtual SDN -enabled LTE
EPC architecture: A case study for S-/P-gateways functions,'' in Proc. IEEE SDN Future Netw. Services
[15] – Rui Wang ; Shanghai Res. Center for Wireless Commun., Shanghai Inst. of Microsyst. & Inf.
Technol., Shanghai, China ; Honglin Hu ; Xiumei Yang. Potentials and Challenges of C -RAN Supporting
Multi -RATs Toward 5G Mobile Networks
[16] J. Bartelt, G. Fettweis, D. Wubben, M. Boldi, and B. Melis, „Heteroge neous backhaul for cloud –
based mobile networks,'' in Proc. IEEE 78th Veh. Technol. Conf., Sep.
[17] – D. Sabella et al., „RAN as a service: Challenges of designing a  flexible RAN architecture in a cloud –
based heterogeneous mobile network,'' in Proc. Future Netw. Mobile Summit, Jul. 2013
[18] – Ericsson. (Feb. 2011). AIR: Antenna Integrated Radio Unit. [Online]. Available:
http://hugin.info/1061 /R/1486615/421909.pdf
[19] – M. Peng, Y. Li, T. Q. S. Quek, and C. Wang, „Device -to-device underlaid cellular networks under
Rician fading channels,'' IEEE Trans. Wireless Commun.
[20] – C. J. Bernardos, A. De Domenico, J. Ortin, P. Rost, and D. Wubben, „Challenges of designing jointly
the backhaul and radio access network in a cloud -based mobile network,'' in Proc. Future Netw. Mobile
Summit, Jul. 2013
[21] – Y. Zhou and W. Yu, „Optimized backhaul compression for uplink cloud radio access network,'' IE EE
J. Sel. Areas Commun.