Context-Aware Distributed Service
Provisioning based on Anycast for
Information-Centric Network
Sha Yuan
∗
, Ding Tang
∗
, Yinlong Liu
∗
, Shuotian Bai
†
, Tao Lin
∗§
and Song Ci
∗‡
∗
High Performance Network Lab, Institute of Acoustics, Chinese Academy of Sciences, Beijing 100190, China
†
School of Computer and Control Engineer, University of Chinese Academy of Sciences, Beijing, China
‡
Department of Computer and Electronics Engineering, University of Nebraska-Lincoln, NE 68182, USA
§
Email: {yuans,tangd,liuyl,lint,sci}@hpnl.ac.cn,
†
baishutian10@mails.ucas.ac.cn
Abstract—Information-Centric Networking (ICN) is a
new emerging concept, in which the principal paradigm
shifts from the traditional end-to-end connection to the
information-centric communication model. In ICN, infor-
mation unit, such as content or service, is distributed
in different sites or data centers to provide large-scale
services. A common supporting approach for scalable
service provisioning is deploying multiple replica servers
throughout the network. Accordingly, an efficient and
flexible scheme is needed to direct distributed requests
to an appropriate replica server. In this paper, we first
propose an incrementally deployable ICN architecture
based on edge/core separation. And then, a practical
anycast-based service provisioning scheme is presented
with joint considerations of both servers contexts and
the underlying network conditions. Extensive experiments
have been performed to evaluate the proposed scheme.
Experiment results show that efficient context-aware dis-
tributed service provisioning can be achieved.
Keywords—Information Centric Network (ICN);
context-awareness; service provisioning; anycast; future
network
I. INTRODUCTION
The communication model of major Internet us-
age has changed from host-centric to information-
centric [1]. However, the TCP/IP mechanism, which
is primarily designed for end-to-end communication,
no longer satisfies the increasing demand for highly
efficient content distribution and service provisioning.
Accordingly, a large volume of duplicated traffic, which
results in the major challenge faced by today’s Internet,
is transmitted in the network. According to the latest
report of Cisco Visual Networking Index, global IP
traffic has increased more than fourfold in the past
five years, and will increase threefold over the next
five years. The annual global IP traffic will surpass the
zettabyte threshold by the end of 2017. Recognizing
this, numerous research activities have been conducted
on the research of Future Internet, such as DONA [2],
PSIRP [3], CCN [4]/NDN [5], SCAFFOLD [6] and
§
Correspondence: No.21 West Road, North 4th Ring Road, Haidian
District, Beijing, China. E-mail: lint@hpnl.ac.cn.
so on. Although these Future Internet research projects
differ with respect to their specific architectures, they
have the concept of Information-Centric Networking
(ICN) in common.
Through in-network caching [7], ICN provides an
efficient communication paradigm to support content-
based applications (e.g., Youtube [8], Hulu), and
service-based applications (e.g., search engine, on-line
gaming and e-commerce). Even so, we argue that repli-
ca servers of these applications will also be needed to
increase the service availability and fault tolerance in
future ICN. Especially for service-based applications,
replica servers are indispensable to support large scale
service. Therefore, an efficient and flexible scheme is
essential for ICN to direct distributed requests to an
appropriate replica to make users as satisfied as possible
in the process of distributed service provisioning. The
focus of this paper is put on this issue for which a novel
anycast-based server selection algorithm is proposed
with joint considerations of both servers contexts and
the underlying network conditions.
The insight of ICN is that users intention is to fetch
content rather than connect to a specific host [9]. In
ICN, uniquely named content is regarded as a core
Internet principle and every network router acts as a
content cache. By decoupling names from locations,
ICN has several benefits, such as lower path stretch and
higher cache hit probability [10]. However, supporting
ICN as a first-order primitive requires all end hosts
and network routers to be replaced. In this way, the
benefits are hugely outweighed by the costs [11]. Since
the wholesale change is unworthy, we present an in-
crementally deployable ICN architecture following the
edge/core separation principle. The whole network is
divided into edge and core. In the edge network, routing
is based on service names, whereas in the core network,
routing is based on locators as todays IP network
does. Indeed, routing scalability caused by billions of
flexible-length service names is one major concern of
the feasibility of ICNs large-scale deployment [12]. In
978-1-4799-3572-7/14/$31.00 ©2014 IEEE