Science China Information Sciences

, Volume 53, Issue 11, pp 2169–2184

A multiple user sharing behaviors based approach for fake file detection in P2P environments

  • Jing Jiang
  • YongJun Li
  • QinYuan Feng
  • Peng Huang
  • YaFei Dai
Research Papers

DOI: 10.1007/s11432-010-4087-5

Cite this article as:
Jiang, J., Li, Y., Feng, Q. et al. Sci. China Inf. Sci. (2010) 53: 2169. doi:10.1007/s11432-010-4087-5
  • 57 Downloads

Abstract

File pollution is a problem that is threatening security and availability in peer-to-peer environments, and could eventually be fatal to the survival of these systems. Current methods of fake file detection, which are based on trust mechanisms or file characteristics, do not consider the problems such as interference of multiple user sharing behaviors, the difficulty in obtaining large amounts of raw data, and malicious fraud immediately after version publication. This is the first time that differences in the sharing habits of users and particularly long retention time for individual files have been detected, which impact fake file detection. This paper proposes a pollution prevention model based on multiple user sharing behaviors, which reduces interference in multiple user sharing behaviors and malicious fraud immediately after version publication. It designs a low-cost implementation mechanism for use in structured peer-to-peer networks, which automatically collects a large amount of user file-sharing information and alleviates difficulties in obtaining large quantities of raw data. A parameter configuration is also given in this paper. Simulation experiments use realistic system logs and prove that this approach detects fake files accurately and quickly, thus decreasing download times of fake files, effectively guaranteeing almost 100% real file downloads and resisting file pollution.

Keywords

peer-to-peer file pollution user sharing behavior 

Copyright information

© Science China Press and Springer-Verlag Berlin Heidelberg 2010

Authors and Affiliations

  • Jing Jiang
    • 1
  • YongJun Li
    • 1
  • QinYuan Feng
    • 1
  • Peng Huang
    • 1
  • YaFei Dai
    • 1
  1. 1.Department of Computer Science and TechnologyPeking UniversityBeijingChina

Personalised recommendations