An analysis of user behavior in a private BitTorrent community

Qing Chao Cai, Kwok Tung Lo

Research output: Journal article publicationJournal articleAcademic researchpeer-review

1 Citation (Scopus)

Abstract

This paper aims to develop an understanding on how users of private BitTorrent communities participate in downloading and uploading by using the download records of CHDBits, one of the largest private BitTorrent communities in China. First, we investigate the arrivals of CHDBits users, which correspond to the initiations of new download tasks. The investigation demonstrates that user arrival rate varies with time in a day, and this variation dovetails well into users' daily schedule. Moreover, we observe that it is with more than half possibility that the two consecutive arrivals of a CHDBits user take place in the same day, and that an individual user issues new download tasks with a probability of 0.23, on which we base to estimate that all CHDBits users are expected to launch at least one new download task in 42 days. Second, we present an analysis on the post-download behavior of users. We observe that the average per task seeding time of 99% of users is no less than 8 h, showing CHDBits users' great willingness to contribute.We also find that almost all the seeding processes have been interrupted and resumed multiple times. Finally, we study the bandwidth distribution of users and evaluate users' bandwidth utilization in the downloading process. Our study shows that the bandwidths of most CHDBits users are uniformly distributed in the range from 256 kB/s to 16 MB/s, and users with a higher bandwidth tend to experience a lower bandwidth utilization.
Original languageEnglish
Pages (from-to)1572-1581
Number of pages10
JournalInternational Journal of Communication Systems
Volume27
Issue number10
DOIs
Publication statusPublished - 1 Jan 2014

Keywords

  • Peer-to-peer
  • Performance analysis
  • Private BitTorrent community
  • Quality of service
  • User behavior

ASJC Scopus subject areas

  • Computer Networks and Communications
  • Electrical and Electronic Engineering

Fingerprint

Dive into the research topics of 'An analysis of user behavior in a private BitTorrent community'. Together they form a unique fingerprint.

Cite this