Improving Performance of ALM Systems with Bayesian Estimation of Peers Dynamics

  • Authors:
  • Ihsan Ullah;Grégory Bonnet;Guillaume Doyen;Dominique Gaïti

  • Affiliations:
  • ERA/Institut Charles Delaunay --- FRE CNRS 2848, Université de Technologie de Troyes, TROYES, France 10000;ERA/Institut Charles Delaunay --- FRE CNRS 2848, Université de Technologie de Troyes, TROYES, France 10000;ERA/Institut Charles Delaunay --- FRE CNRS 2848, Université de Technologie de Troyes, TROYES, France 10000;ERA/Institut Charles Delaunay --- FRE CNRS 2848, Université de Technologie de Troyes, TROYES, France 10000

  • Venue:
  • MMNS 2009 Proceedings of the 12th IFIP/IEEE International Conference on Management of Multimedia and Mobile Networks and Services: Wired-Wireless Multimedia Networks and Services Management
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

P2P-based Application Layer Multicast (ALM) systems have shown a great success for several group communication applications. But some performance problems still await a major breakthrough from these systems for critical services such as live video streaming. For these applications, one of the problems is the dynamics of users' presence since the unannounced departure of a peer causes an interruption in service for all dependent ones. In this paper, we address this issue and propose a probabilistic approach based on Bayesian inference to anticipate users' departures and let peers react proactively. Through simulations and experimental evaluation, we prove that our approach improves significantly the performance of ALM systems with a low overhead.