中国邮电高校学报(英文) ›› 2010, Vol. 17 ›› Issue (2): 85-91.doi: 10.1016/S1005-8885(09)60452-0

• Networks • 上一篇    下一篇

Relaying Strategy for Peer-to-Peer Content Distribution Based On Genetic Algorithm

何倩1,陈俊亮2,孟祥武2,商彦磊2   

  1. 1. 北京邮电大学
    2. 北京邮电大学网络与交换技术国家重点实验室
  • 收稿日期:2009-05-25 修回日期:2009-09-29 出版日期:2010-04-30 发布日期:2010-06-01
  • 通讯作者: 何倩 E-mail:treeqian@gmail.com
  • 基金资助:

    国家级.国家自然科学基金;国家级.国家“863计划”项目

Relaying Strategy for Peer-to-Peer Content Distribution Based On Genetic Algorithm

  • Received:2009-05-25 Revised:2009-09-29 Online:2010-04-30 Published:2010-06-01

摘要:

This article proposes a cooperative relaying strategy to make efficient use of the relaying resources of Internet Service Providers (ISPs), speedup distribution and save server bandwidth costs. ISPs cooperatively relay for each other, and peers assist to distribute and fetch the content as near as possible. Base on fluid model, a constrained model is derived to get optimized global distribution performance in the channel-based system with limited relaying resources. The multi-objectives of the model are minimizing the average distribution time and the distribution time of the slowest channel. Genetic Algorithm (GA) is designed to solve the optimization problem. The relaying strategy based on GA can be run periodically to update the allocation policy of ISPs. The relaying strategy’s distribution performance is analyzed in the experiments and results show that GA can give proper solutions for various network topologies.

关键词:

Resource Allocation

Abstract:

This article proposes a cooperative relaying strategy to make efficient use of the relaying resources of Internet Service Providers (ISPs), speedup distribution and save server bandwidth costs. ISPs cooperatively relay for each other, and peers assist to distribute and fetch the content as near as possible. Base on fluid model, a constrained model is derived to get optimized global distribution performance in the channel-based system with limited relaying resources. The multi-objectives of the model are minimizing the average distribution time and the distribution time of the slowest channel. Genetic Algorithm (GA) is designed to solve the optimization problem. The relaying strategy based on GA can be run periodically to update the allocation policy of ISPs. The relaying strategy’s distribution performance is analyzed in the experiments and results show that GA can give proper solutions for various network topologies.

Key words:

Resource Allocation