Optimizing peer selection in BitTorrent networks with genetic algorithms

  • Authors:
  • Tiejun Wu;Maozhen Li;Man Qi

  • Affiliations:
  • School of Engineering and Design, Brunel University, Uxbridge, UB8 3PH, UK;School of Engineering and Design, Brunel University, Uxbridge, UB8 3PH, UK;Department of Computing, Canterbury Christ Church University, Canterbury, Kent, CT1 1QU, UK

  • Venue:
  • Future Generation Computer Systems
  • Year:
  • 2010

Quantified Score

Hi-index 0.00

Visualization

Abstract

BitTorrent has emerged as an effective peer-to-peer application for digital content distribution in the Internet. However, selecting peers in BitTorrent for efficient content distribution still poses a number of challenges due to high heterogeneities of peers with varied rates of uploading bandwidth and dynamic content. This paper presents GA-BT, a genetic algorithm based peer selection optimization strategy for efficient content distribution in BitTorrent networks taking into account both the uploading bandwidth of peers and the availability of content among peers. GA-BT employs the divisible load theory to dynamically predict optimal fitness values to speed up the convergence process in producing optimal or near optimal solutions in peer selection. A BitTorrent simulator is implemented for GA-BT performance evaluation, and the experimental results show the effectiveness of GA-BT in peer selection optimization.