Architecture and performance models for QoS-driven effective peering of content delivery networks

  • Authors:
  • Mukaddim Pathan;Rajkumar Buyya

  • Affiliations:
  • (Correspd. E-mail: apathan@csse.unimelb.edu.au) Grid Computing and Distributed Systems (GRIDS) Laboratory, Department of Computer Science and Software Engineering, The University of Melbourne, Par ...;Grid Computing and Distributed Systems (GRIDS) Laboratory, Department of Computer Science and Software Engineering, The University of Melbourne, Parkville, VIC 3010, Australia

  • Venue:
  • Multiagent and Grid Systems - Content management and delivery through P2P-based content networks
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

The proprietary nature of existing Content Delivery Networks (CDNs) means they are closed and do not naturally cooperate. A CDN is expected to provide high performance Internet content delivery through global coverage, which might be an obstacle for new CDN providers, as well as affecting commercial viability of existing ones. Finding ways for distinct CDNs to coordinate and cooperate with other CDNs is necessary to achieve better overall service, as perceived by end-users, at lower cost. In this paper, we present an architecture to support peering arrangements between CDNs, based on a Virtual Organization (VO) model. Our approach promotes peering among providers, while upholding user perceived performance. This is achieved through proper policy management of negotiated Service Level Agreements (SLAs) between peers. We also present a Quality of Service (QoS)-driven performance modeling approach for peering CDNs in order to predict the user perceived performance. We show that peering between CDNs upholds user perceived performance by satisfying the target QoS. The methodology presented in this paper provides CDNs a way to dynamically distribute user requests to other peers according to different request-redirection policies. The model-based approach helps an overloaded CDN to return to a normal state by offloading excess requests to the peers. It also assists in making concrete QoS guarantee for a CDN provider. Our approach endeavors to achieve scalability and resource sharing among CDNs through effective peering in a user transparent manner, thus evolving past the current landscape where non-cooperative and distinct} CDNs exist.