Page 141 - ITU Kaleidoscope 2016

A POPULARITY BASED CACHING STRATEGY FOR THE FUTURE INTERNET

Suhaidi Hassan, Ikram Ud Din, Adib Habbal, and Nur Haryani Zakaria

InterNetWorks Research Laboratory, School of Computing, Universiti Utara Malaysia

ABSTRACT It is therefore important to deduce that the existing Internet
architecture can be considered as a restricting factor of the
Information-Centric Networking (ICN) is an attractive net- current Internet growth and the design of new applications.
work model receiving increasing consideration by the re-
To add more, several studies, for example, [5,6] have shown
search community because of its inspiring features. To better that the advancement of the Internet design is driven by in-
manage the Internet usage move from host-centric communi-
cremental and responsive increases and therefore change in
cation to receiver-driven content retrieval, revolutionary ICN
the architecture is indispensable. In this manner, globally
architectures have been proposed. A distinguished character- the research community is working towards understanding
istic of these innovative architectures is to provide ubiquitous
the architectural challenges to decide the rule that will drive
and transparent in-network caching to enhance network re-
the future Internet design. Numerous research projects (e.g.,
source utilization and accelerate content dissemination. With US NSF GENI [7], EU-FIA [8], and AsiaFI [9]) have been
the exponential increase of Internet traffic, the issue of con-
funded in the last few years to define the existing limitations
tent storage is a growing concern in ICN. In this paper, we and future needs for the Internet architecture, and therefore
present a caching strategy that considerably increases cache
Information-Centric Networking (ICN) [1,10–12]] is one of
hit rate and reduces stretch ratio, which are the most im-
the considerable results of these research activities. ICN has
portant metrics in the evaluation of ICN caching. Through different modules, such as naming, routing, mobility, security,
extensive simulations, it is shown that our proposed work is a
and caching. However, the majority of our research commu-
favorable and realistic contribution for the standardization
nity is attracted by the caching module because of the limited
exercise of data caching for achieving accurate and valid cache size of network nodes. In ICN caching, the network
network performance in the future Internet.
nodes have the ability to cache a content locally once it is
Keywords— Future Internet, ICN, CPCE, caching, content downloaded by the end users. Therefore, if new requests ar-
popularity rive for the same content, they are satisfied locally rather than
contacting the original server.
1. INTRODUCTION
2. PROBLEM DEFINITION
Toward the start of the Internet, clients were scholastic in
nature, for the most part inspired by mail trade and document In ICN, the contents are cached locally by network nodes
exchanges [1]. Moreover, sharing of resources was a vital
(e.g., routers). These routers may pose strict constraints with
issue that forced significant difficulties with respect to cor- respect to cache management. Thus management becomes
respondence among end hosts [2]. However, since the last
a critical issue for content caching and caching strategies.
decade the Internet popularity has brought about the activity When the network becomes stable and the router’s cache over-
on the Internet to become significantly reliable [3]. Infor- flows, a replacement policy, such as Least Recently Used
mation sharing and dissemination, for example, scholastic,
(LRU), Least Frequently Used (LFU), or Random policy is
social, and business, over the Internet is the major cause of the used to evict one of the cached contents to make room for the
Internet growth. Distribution of named data is a noteworthy new arrived one. However, besides the replacement policy,
application in the existing Internet. Along with online content
content caching is the main issue, i.e., which content should
dissemination, other distribution technologies, for instance, be cached and at which location it needs to be stored so that
Content Distribution Network (CDN) and Peer-to-Peer (P2P)
to efficiently utilize memory and bandwidth consumption.
communication have been well developed and are advancing For that, many strategies have been proposed, for example,
the communication framework for getting contents by name, Cache Everything Everywhere (CEE) [2] - the default ICN
regardless of the location of the main server [3,4]. Keeping in
strategy, Cache Less for More [13] – caches contents at a node
mind the end goal to react to expanding activity volume in the which has the maximum betweenness centrality, Probabilistic
existing Internet for applications are utilized that use caching Caching (ProbCache) [14] – stores contents near the users,
and content distribution and replication in various particular
Cooperative In-network Caching (CIC) [15] - where the con-
ways [1]. tent is divided into different chunks and cached at more than
one node, Cache Aware Target Identification (CATT) [16,17]-
This research is supported by the Ministry of Higher Education Malaysia
Research Grant Scheme with reference number FRGS/UUM/13258. where the content is cached at a single node on the publisher-

978-92-61-20431-0/CFP1668P-ART © 2016 ITU – 123 – Kaleidoscope