网络虚拟化_网络虚拟化资源-CSDN文库

需积分: 10 180 浏览量 2015-04-02 09:36:33 上传评论收藏 559KB PDF 举报

资源推荐

资源详情

资源评论

http://www.ieee.org/publications_standards/publications/rights/index.html for more information.

This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI

10.1109/COMST.2014.2352118, IEEE Communications Surveys & Tutorials

Wireless Network Virtualization: A Survey, Some

Research Issues and Challenges

Chengchao Liang and F. Richard Yu

∗

Depart. of Systems and Computer Eng., Carleton University, Ottawa, ON, Canada

Email: chengchaoliang@sce.carleton.ca; richard.yu@carleton.ca

Abstract—Since wireless network virtualization enables ab-

straction and sharing of infrastructure and radio spectrum

resources, the overall expenses of wireless network deployment

and operation can be reduced signiﬁcantly. Moreover, wireless

network virtualization can provide easier migration to newer

products or technologies by isolating part of the network.

Despite the potential vision of wireless network virtualization,

several signiﬁcant research challenges remain to be addressed

before widespread deployment of wireless network virtualiza-

tion, including isolation, control signaling, resource discovery

and allocation, mobility management, network management and

operation, and security as well as non-technical issues such as

governance regulations, etc. In this paper, we provide a brief

survey on some of the works that have already been done

to achieve wireless network virtualization, and discuss some

research issues and challenges. We identify several important

aspects of wireless network virtualization: overview, motivations,

framework, performance metrics, enabling technologies, and

challenges. Finally, we explore some broader perspectives in

realizing wireless network virtualization.

Index Terms—Wireless network virtualization, abstraction and

sharing, isolation, cognitive radio and networks, cloud computing.

I. INTRODUCTION

In the information and communications technology (ICT)

sector, virtualization has become a popular concept in dif-

ferent areas, e.g., virtual memory [1], virtual machines [2],

virtual storage access network [3] and virtual data centers [4].

Virtualization involves abstraction and sharing of resources

among different parties. With virtualization, the overall cost of

equipment and management can be signiﬁcantly reduced due

to the increased hardware utilization, decoupled functionalities

from infrastructure, easier migration to newer services and

products, and ﬂexible management [1]–[4].

In wired networks, virtualization has occurred for decades,

e.g., virtual private networks (VPNs) over wide area networks

(WANs) and virtual local area networks (VLANs) in enterprise

networks [5], [6]. Recently, network virtualization has been ac-

tively used in Internet research testbeds, such as G-Lab [7] and

4WARD [8], and applied in the cloud computing environment

[9]. It aims to overcome the resistance of the current Internet to

fundamental architecture changes. Network virtualization has

been considered as one of the most promising technologies

for the future Internet [10].

With the tremendous growth in wireless trafﬁc and services,

it is natural to extend virtualization to wireless networks.

With wireless network virtualization, network infrastructure

can be decoupled from the services that it provides, where

differentiated services can coexist on the same infrastructure,

maximizing its utilization [11]. Consequently, multiple wire-

less virtual networks operated by different service providers

(SPs) can dynamically share the physical substrate wireless

networks operated by mobile network operators (MNOs).

Since wireless network virtualization enables the sharing of in-

frastructure and radio spectrum resources, the capital expenses

(CapEx) and operation expenses (OpEx) of wireless (radio)

access networks (RANs), as well as core networks (CNs),

can be reduced signiﬁcantly. Moreover, mobile virtual network

operators (MVNOs) who may provide some speciﬁc telecom

services (e.g., VoIP, video call, over-the-top services) can help

MNOs attract more users, while MNOs can produce more

revenue by leasing the isolated virtualized networks to them

and evaluating some new services [12]. Meanwhile, wireless

network virtualization provides easier migration to newer

products or technologies while supporting legacy products

by isolating part of the network [12], [13]. In addition, the

emerging heterogeneous wireless networks need a convergent

and powerful network management mechanism, which can be

provided by wireless network virtualization [14].

Despite the potential vision of wireless network virtual-

ization, several signiﬁcant research challenges remain to be

addressed before widespread deployment of wireless network

virtualization, including isolation, control signaling, resource

discovery and allocation, mobility management, network man-

agement and operation, and security as well as non-technical

issues such as governance regulations, etc. Particularly, unlike

wired networks, where bandwidth resource abstraction and

isolation can be done on a hardware (e.g., port and link)

basis, radio resource abstraction and isolation is not straight-

forward, due to the inherent broadcast nature of wireless

communications and stochastic ﬂuctuation of wireless channel

quality. Another signiﬁcant challenge of wireless network

virtualization is resource allocation, which decides how to

embed a virtual wireless network on physical networks. In

addition, a large number of intelligent devices/nodes with self

adaptation/context awareness capabilities induce non-trivial

security challenges to wireless network virtualization. These

challenges need to be tackled broadly by comprehensive

research effort.

In this paper, we provide a brief survey on some of

the works that have already been done to achieve wireless

network virtualization, and discuss some research issues and

challenges. A taxonomy graph of our approach towards the

http://www.ieee.org/publications_standards/publications/rights/index.html for more information.

This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI

10.1109/COMST.2014.2352118, IEEE Communications Surveys & Tutorials

Motivations

Wireless Network

Virtualization

Business Models

Requirements

Overview

Brief History

SDN and OpenFlow

Projects

Performance Metrics

Traditional Me trics

Virtualization-Specific Metrics

Framework

Radio Spectrum Resource

Wi rele ss Ne twork I n frastructure

Wireless Virtual Resource

Enabling Technologies

IEEE 802.11-based

Cellular-based

IEEE 802.16-based

Challenges

Isolation

Control Signaling

Resour c e Dis covery and

Allocation

Mo bi lity Manage me nt

Wireless Virtualization

Controller

Others

Network Management

Security

Broader Perspectives

Cognitive Radio and Networks

Multi-tier Wireless Networks

Cloud Computing

Fig. 1. Road map of wireless network virtualization.

design of wireless network virtualization is given in Fig. 1.

As shown in the ﬁgure, we identify seven important aspects

of wireless network virtualization where we would like to fo-

cus: overview, motivations, framework, performance metrics,

enabling technologies, challenges, and broader perspectives.

In the following sections, we elaborate on each such aspect

and discuss the related issues. Section II presents a brief

history of wired network virtualization. Then some projects

on network virtualization are introduced. Software deﬁned

networking and OpenFlow are also presented. In Section

III, we will present the business models and the involved

parties. The motivations and requirements of wireless network

virtualization are also discussed. A framework is summarized

in Section IV with four main components: radio spectrum

resource, wireless network infrastructure, wireless virtual re-

source, and wireless virtualization control. Section V presents

some performance metrics that are necessary to evaluate the

performance and quality of a virtualized wireless networks.

Some enabling technologies for wireless network virtualization

are discussed in Section VI according to different radio access

technologies. Section VII presents research issues and chal-

lenges. Some broader perspectives are also presented. Finally,

we conclude this study in Section VIII.

II. O

VERVIEW OF NETWORK VIRTUALIZATION

In this section, we ﬁrst present a brief history of wired

network virtualization. Then, some projects on network vir-

tualization are presented. Software deﬁned networking (SDN)

and OpenFlow are also introduced in this section.

A. Brief History of Wired Network Virtualization

Virtualization has occurred in wired networks for decades.

Some examples of wired network virtualization include vir-

tual local area networks (VLANs), virtual private networks

(VPNs), active and programmable networks, and overlay net-

works, which are described in the following.

1) VLAN: A VLAN refers to a domain where a group

of hosts with a common interest are allowed to be logically

brought together under a single broadcast domain regardless

of their physical connectivity [15].

2) VPN: In a VPN, a private network, whose hosts are

distributed in multiple sites, connects through private and

secured tunnels (links) over public communication networks

(e.g., the Internet or PSDN) [16]–[18]. Depended on different

layers, VPNs can be classiﬁed into four classes: Layer 1

VPNs [19], Layer 2 VPNs [20], Layer 3 VPNs [21] and

higher layer VPNs [22]. It should be noted that Layer 1 VPNs

have no guarantee between data plane connectivity and control

http://www.ieee.org/publications_standards/publications/rights/index.html for more information.

This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI

10.1109/COMST.2014.2352118, IEEE Communications Surveys & Tutorials

plane connectivity, which means that each service network has

independent address space and L1 resource view, separating

policies and completing isolation from other VPNs. This is the

main difference between Layer 1 VPNs and Layer 2/3 VPNs.

3) Active and programmable networks: In response to

user demands, the need of creating, deploying, and manag-

ing novel services on the ﬂy drives the research of active

and programmable networks. The fundamental discussion of

programmable networks is that separating communications

hardware from control software and allowing multiple parties

to run possibly totally different protocols on the same network

elements without any conﬂict. The open signaling approach

and the active networks approach are the two kinds of imple-

mentations of active and programmable networks [23].

4) Overlay network: A virtual network that creates a virtual

topology based on the physical topology of another network

can be considered as an overlay network where nodes are

connected through virtual links, which correspond to paths in

the underlying network. Overlays are typically implemented

in the application layer [6].

In the above four examples of wired network virtualization,

the scope of virtualization is limited to one or two layers.

However, to exploit the full beneﬁts of virtualization, the

network needs to be fully virtualized, and services are clearly

separated from their underlying infrastructure.

B. Projects on Network Virtualization

Recently, several research projects have been started around

the world in the area of network virtualization, including X-

Bone [24] and Tempes [25] focusing on networking technol-

ogy; UCLP [26], VNET [27], AGAVE [28] and VIOLIN [29]

focusing on layers of virtualization; VNRMS [30], NetScript

[31], Genesis [32] and FEDERICA [33] focusing on archi-

tectural domain and management; and PlanetLab [34], GENI

[35], VINI [36], CABO [37], 4WARD [8] and NouVeau [38]

focusing on the granularity of virtualization; and VITRO [39]

focusing on virtualization of wireless sensor networks. Due

to the space limitation, we only give a brief introduction to

CABO, GENI, 4WARD and PlanetLab, which are important

projects on network virtualization.

1) CABO: In CABO, the concept of separation between

infrastructure providers (InPs) and SPs is promoted and im-

proved by an integrated project to support full virtualization

that allows SPs to provide end-to-end services over multiple

InPs’ infrastructure. CABO is also the ﬁrst full virtualization

project in which virtual routers can move (are mapped) from

one physical node to another. It also provides guarantees

and customization to service providers to support end-to-end

services to the end users.

2) 4WARD: In 4WARD, more detailed business models

are introduced in addition to InPs and SPs, including vir-

tual network providers (VNPs) and virtual network operators

(VNOs). This business model gives more opportunities to

the market. The project also includes substantial work on

resource allocation and resource discovery of network virtu-

alization. Moreover, 4WARD also supports virtualization of

heterogeneous networking technologies. Another signiﬁcant

contribution is that 4WARD implements network virtualization

not only in experimental networks and testbeds but also in

realistic networks.

3) PlanetLab: PlanetLab proposes a concept of slice-

ability, in which each application acquires and runs in a

slice of the overlay. Slice-ability is a crucial ability and

design principle in network virtualization, which dominates the

realization of both wired and wireless network virtualization.

4) GENI: GENI introduces network virtualization to the

wireless area. In GENI, virtualization techniques and slicing

techniques are proposed by utilizing TDMA, FDMA and

SDMA. Moreover, GENI gives researchers the opportunity to

create customized virtual networks unfettered by assumptions

or requirements of the existing Internet.

5) VITRO: VITRO proposes an integrated architecture of

enabling virtualization in wireless sensor networks while pro-

viding advanced services. The approach in VITRO realizes

the decoupling of the applications running on physical nodes

from the physical sensor deployment. This concept of virtual

sensor networking allows the dynamic cooperation among

sensor nodes, helping the proliferation of new services and

applications beyond the scope of the original deployment.

C. Software Deﬁned Networking and OpenFlow

SDN is an emerging network architecture where network

control is decoupled from forwarding and is directly pro-

grammable [40]. It is considered as one of the most promising

technologies to realize virtual networks, especially in network

control. SDN focuses on four key features [41]:

• Separation of the control plane from the data plane.

• A centralized controller and view of the network.

• Open interfaces between the devices in the control plane

(controllers) and those in the data plane.

• Programmability of the network by external applications.

By separating a network’s control logic from the underlying

physical routers and switches that forward trafﬁc, network

operators can write high-level control programs that specify

the behavior of an entire network. This is different from

conventional networks, where network operators must codify

functionalities in terms of low-level device conﬁgurations.

SDN allows network administrators to have programmable

central control of network trafﬁc via a controller without

requiring physical access to the network’s switches. A conﬁgu-

ration of SDN can create a logical network control plane where

hardware is physically decoupled from the data forwarding

plane hardware, i.e., a network switch can forward packets

and a separate server can run the network control plane. The

decoupling allows for the control plane to be implemented

using a different distribution model than the data plane.

Control plane development and runtime environment tasks can

then be run on a different platform (other than the low-powered

management CPUs found on hardware switches and routers).

OpenFlow is a standard communications interface deﬁned

between the control and forwarding layers of an SDN archi-

tecture [42]. The standard is managed by Open Networking

Foundation (ONF). OpenFlow allows direct access to and

manipulation of the forwarding plane of network devices, such

http://www.ieee.org/publications_standards/publications/rights/index.html for more information.

This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI

10.1109/COMST.2014.2352118, IEEE Communications Surveys & Tutorials

as switches and routers. With OpenFlow, the path of network

packets through the network of switches can be determined by

software running on multiple routers. A number of network

switch and router vendors have announced intent to support

OpenFlow standard.

III. M

OTIVATIONS,BUSINESS MODELS &REQUIREMENTS

WIRELESS NETWORK VIRTUALIZATION

In this section, we will discuss the motivations of wireless

network virtualization. Business models with different roles in

the wireless network market and the functions of these roles

will be presented. Moreover, we will discuss the requirements

that need to be met to implement wireless network virtualiza-

tion.

A. What is Wireless Network Virtualization?

Wireless network virtualization can have a very broad scope

ranging from spectrum sharing, infrastructure virtualization, to

air interface virtualization. Similar to wired network virtual-

ization, in which physical infrastructure owned by one or more

providers can be shared among multiple service providers,

wireless network virtualization needs the physical wireless

infrastructure and radio resources to be abstracted and isolated

to a number of virtual resources, which then can be offered

to different service providers. In other words, virtualization,

regardless of wired or wireless networks, can be considered

as a process splitting the entire network system [13]. However,

the distinctive properties of the wireless environment, in terms

of time-various channels, attenuation, mobility, broadcast, etc.,

make the problem more complicated. Furthermore, wireless

network virtualization depends on speciﬁc access technologies,

and wireless network contains much more access technologies

compared to wired network virtualization and each access

technology has its particular characteristics, which makes

convergence, sharing and abstraction difﬁcult to achieve.

Therefore, it may be inaccurate to consider wireless network

virtualization as a subset of network virtualization [11].

In this paper, we consider wireless network virtualization as

the technologies in which physical wireless network infrastruc-

ture resources and physical radio resources can be abstracted

and sliced into virtual wireless network resources holding

certain corresponding functionalities, and shared by multiple

parties through isolating each other. In other words, virtualiz-

ing wireless network is to realize the process of abstracting,

slicing, isolating and sharing the wireless networks. Since

wireless network resources are sliced into multiple slices, the

terms of virtual slice and virtual network have the similar

meanings of virtual wireless network resources. We may use

them alternatively in this paper

B. What are the Business Models of Wireless Network Virtu-

alization?

In wireless network virtualization, physical resources are

owned by some parties, and virtual resources are utilized by

some other parties. The question is who these parties are.

Business models can describe the constitution of the roles in

the wireless network market and the main functions of these

roles. As shown in Fig. 2(a), generally, after wireless network

virtualization, there are two logical roles, MNO and SP [12],

[37], [43]. All of the infrastructures and radio resources of

physical substrate wireless networks, including the licensed

spectrum, radio access networks (RANs), backhaul, transmis-

sion networks (TNs), and core networks (CNs), are owned and

operated by MNOs. MNOs execute the virtualization of the

physical substrate networks into some virtual wireless network

resources. For brevity, we use virtual resources to indicate

the virtual wireless network resources. SPs lease these virtual

resources, operate and program them so that to offer end-to-

end services to end users. In some papers (e.g., [5]), the MNO

becomes InP, which is only responsible for owing and leasing

wireless network resources to SPs. SPs will create and deploy

the virtual resource by themselves based on the leased and

allocated resource to satisfy the requirements of end-to-end

services.

The roles in in the above business models can be further

decoupled into more specialized roles, including InP, mobile

virtual network provider (MVNP), mobile virtual network

operator (MVNO) and SP [44]–[46], as shown in Fig. 2(b).

The functions of them are describing as follows.

a) InP: owns the infrastructure and wireless network

resources. In some cases, the spectrum resources may or may

not be owned by InP.

b) MVNP: leases the network resources and creates

virtual resources. Some MVNPs may have some licensed spec-

trum such that they do not need request spectrum resources

from InP. In some papers (e.g., [47]), MVNP is called mobile

virtual network enabler (MVNE).

c) MVNO: operates and assigns the virtual resources to

SPs. Meanwhile, in some approaches, MVNOs consists of

the roles of both MVNOs and MVNPs. Actually, this model

is ﬁt for the emerging concept of so called XaaS [48] in

cloud computing. What provided in InPs is infrastructure-as-

a-service (IaaS) while what provided in MVNOs is network-

as-a-service (NaaS).

d) SP: concentrates on providing services to its sub-

scribers based on the virtual resources provided by MVNOs.

In other words, virtual resources are requested by SPs,

managed by MVNOs, created by MVNPs, and running at InPs

physically. Obviously, this four-level model can create more

opportunities in the market and simplify the functions of each

role intuitively. Nevertheless, more coordination mechanisms

and interfaces should be used, which may increase the com-

plexity and latency signiﬁcantly.

Here, we give a short discussion on the role of MVNOs.

The deﬁnition of MVNOs is different in different countries

and communities [47]. The authors of [13] argue that MVNOs

do not own any spectrum and radio access networks for its

subscribers to access. However, in [47], MVNOs may or may

not own infrastructure, or may only own other parts of wireless

networks (e.g., CNs) except RANs and spectrum licenses. The

authors of [47] consider that MVNOs are the key players

who can break the value chain of telecommunications both

wired and wireless in the future mobile network markets. This

claim is based on the existence of pure InPs, which means

剩余23页未读，继续阅读

评论收藏

内容反馈

280918690

粉丝: 0
资源: 2

网络虚拟化

网络虚拟化技术

什么是网络虚拟化 网络虚拟化简介

VXLAN与园区网络虚拟化

系统虚拟化

虚拟化

虚拟化技术

虚拟化解决方案

虚拟化介绍

虚拟网络创建

网络也走虚拟化道路

网络虚拟化技术介绍.pdf

网络虚拟化技术在云计算领域应用探讨

网络虚拟化概述.pptx

ONS 2014：利用网络虚拟化技术优化网络性能

虚拟专用网络

基于云计算的计算机网络虚拟化实验设计.pdf

技术白皮书-NSX网络虚拟化平台简介

研究生毕业论文网络虚拟化开题报告.doc

网络虚拟软件（路由器，交换机配置软件）

VMware NSX网络虚拟化概览

解决Win 10与不兼容VirtualBox操作过程文档+（附带软件）.zip

计算机网络知识点总结(谢希仁第八版).pdf

最新资源

什么是网络虚拟化网络虚拟化简介