User behaviors in private BitTorrent communities

Adele L. Jia*, Xiaowei Chen, Xiaowen CHU, Johan A. Pouwelse, Dick H.J. Epema

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

3 Citations (Scopus)

Abstract

Many private BitTorrent communities employ Sharing Ratio Enforcement (SRE) schemes to incentivize users to contribute. It has been demonstrated that users in private communities are highly dedicated and that they seed much longer than users in communities where SRE is not employed. While most previous studies focus on showing the positive effect of user dedication in achieving high download speed, in this paper we explore the user behaviors in private communities, we argue the reasons for these behaviors, and we demonstrate both the positive and the negative effects of these behaviors. We show that under SRE, users seed for excessively long times to maintain required sharing ratios, but that their seedings are often not very productive (in terms of low upload speed) and that their long seeding times do not necessarily lead to large upload amounts. We find that as users evolve in the community, some users become more committed, in terms of increasing ratios between their seeding and leeching times. In the mean time, some users game the system by keeping risky and low sharing ratios while leeching more often than seeding. Based on these observations, we analyze strategies that alleviate the negative effects of these user behaviors from both the user's and the community administrator's perspective.

Original languageEnglish
Pages (from-to)34-45
Number of pages12
JournalComputer Networks
Volume60
DOIs
Publication statusPublished - 26 Feb 2014

Scopus Subject Areas

  • Computer Networks and Communications

User-Defined Keywords

  • Incentive policy
  • Peer-to-Peer
  • Private BitTorrent community
  • Sharing ratio enfocement
  • User behavior

Fingerprint

Dive into the research topics of 'User behaviors in private BitTorrent communities'. Together they form a unique fingerprint.

Cite this