An Overview of Cloud Based Content Delivery Networks: Research Dimensions and State-of-the-Art - Transactions on Large-Scale-Data-and Knowledge-Centered Systems XX - page 131

An Overview of Cloud Based Content Delivery
Networks: Research Dimensions
and State-of-the-Art

Meisong Wang 1,2 , Prem Prakash Jayaraman 2 , Rajiv Ranjan 2 ,
Karan Mitra 3 , Miranda Zhang 1,2 , Eddie Li 1,2 , Samee Khan 4 ,
Mukkaddim Pathan 5 , and Dimitrios Georgeakopoulos 6( & )

1 Research School of Computer Science, ANU, Canberra, Australia
{u5454816,miranda.zhang,zheng.li}@anu.edu.au
2 CSIRO DP&S Flagship, Canberra, Australia
{prem.jayaraman,rajiv.ranjan}@csiro.au
3 Lule å University of Technology, SE-931 87 Skellefte å , Sweden
[e-mail]
4 North Dakota State University, Fargo, USA
[e-mail]
5 Telstra Corporation, Melbourne, Australia
[e-mail]
6 Royal Melbourne Institute of Technology, Melbourne, Australia
[e-mail]

Abstract. Content distribution networks (CDNs) using cloud resources such as
storage and compute have started to emerge. Unlike traditional CDNs hosted on
private data centers, cloud-based CDNs take advantage of the geographical
availability and the pay-as-you-go model of cloud platforms. The Cloud-based
CDNs (CCDNs) promote content-delivery-as-a-service cloud model. Though
CDNs and CCDNs share similar functionalities, introduction of cloud impose
additional challenges that have to be addressed for a successful CCDN
deployment. Several papers have tried to address the issues and challenges
around CDN with varying degree of success. However, to the best of our
knowledge there is no clear articulation of issues and challenges problems
within the context of cloud-based CDNs. Hence, this paper aims to identify the
open challenges in cloud-based CDNs. In this regard, we present an overview of
cloud-based CDN followed by a detailed discussion on open challenges and
research dimensions. We present a state-of-the-art survey on current commercial
and research/academic CCDNs. Finally, we present a comprehensive analysis of
current CCDNs against the identied research dimensions.

1 Introduction

The digital universe is doubling in size every two years. It is expected that the data we
create and copy will reach 44 zettabytes by 2020 [ 19 ]. The global internet video traf c

alone will comprise 79 percent of all Internet traf c in 2016, up from 66 percent in 2013

[ 18 ]. In our current Internet-driven world, consumers expect fast, always-on data access
from anywhere and any device. As a result, content providers are expected to confront