Comparing paedophile activity in different P2P systems

Raphaël Fournier, Thibault Cholez, Matthieu Latapy, Clémence Magnien, Isabelle Chrisment, Ivan Daniloff, Olivier FestorarXiv 2012 · OpenReview

Abstract

Peer-to-peer (P2P) systems are widely used to exchange content over the Internet. Knowledge on paedophile activity in such networks remains limited while it has important social consequences. Moreover, though there are different P2P systems in use, previous academic works on this topic focused on one system at a time and their results are not directly comparable. We design a methodology for comparing KAD and eDonkey, two P2P systems among the most prominent ones and with different anonymity levels. We monitor two eDonkey servers and the KAD network during several days and record hundreds of thousands of keyword-based queries. We detect paedophile-related queries with a previously validated tool and we propose, for the first time, a large-scale comparison of paedophile activity in two different P2P systems. We conclude that there are significantly fewer paedophile queries in KAD than in eDonkey (approximately 0.09% vs 0.25%).

Introduction

While P2P file-sharing networks are often associated with the exchange of illegal content, quantifying this activity accurately is a significant challenge. Previous research has typically focused on single networks using disparate methodologies, making it impossible to compare activity levels across different architectures. This study addresses this gap by comparing the prevalence of paedophile activity in two major P2P systems: the server-based eDonkey and the fully decentralized KAD. The central questions are whether network architecture—centralized vs. decentralized—and the level of anonymity influence the prevalence of such activity.

Methodology

To ensure a fair comparison, the researchers developed a normalized monitoring approach across three datasets collected in 2010:

  • eDonkey Monitoring: Queries were recorded from two large servers: one in France (eDonkeyFR) with strict filtering and one in Ukraine (eDonkeyUA) with an open policy.
  • KAD Monitoring: As a decentralized network, KAD was monitored using the HAMACK architecture, which uses distributed probes to capture queries.
  • Normalization: Analysis was restricted to single-keyword queries from a curated list of 72 keywords, categorized into "Paedophile" (unambiguous terms), "Mixed" (primarily age indicators like 12yo), and "Not Paedophile" (control terms like linux).
  • Inference Model: Because KAD data is captured via probing, a probabilistic model was used to infer the total fraction of paedophile queries in the entire network based on observed frequencies in the eDonkey datasets.

Comparative Prevalence

The study found that the estimated fraction of paedophile queries in the total eDonkey network is approximately 0.25%, whereas in KAD it is significantly lower at 0.09%. This finding is supported by the distribution of queries within the monitored subset of 72 keywords.

Distribution of monitored queries across categories for KAD and eDonkey datasets.

As shown in Figure 1, the proportion of "paedophile" category queries within the monitored set is markedly lower in KAD compared to both eDonkey servers. Interestingly, the filtering policy of the eDonkey server (French vs. Ukrainian) had no significant impact on the proportion of these queries, suggesting that server reputation does not strongly influence user behavior in this context.

Behavioral Consistency

Despite the difference in overall volume, the nature of the activity remained remarkably consistent across all monitored networks. This is most evident in the distribution of age indicators within the "mixed" keyword category.

Frequency of age-related keywords across the three datasets.

Figure 4 demonstrates that the frequency of queries for specific ages (e.g., "1yo" to "16yo") follows an almost identical pattern across KAD, eDonkeyFR, and eDonkeyUA, with a clear peak at "12yo". This suggests a consistent target demographic for this activity regardless of the platform's architecture or anonymity level.

Discussion: The Anonymity Paradox

The results present a paradox: KAD, which offers higher perceived anonymity due to its decentralized nature, hosts significantly less paedophile activity than the less-anonymous eDonkey. The authors hypothesize that this may be due to technical barriers. KAD requires more skill to navigate and uses UDP, which is more difficult to tunnel through anonymity tools like Tor compared to the TCP-based eDonkey.

Conclusion

This large-scale comparative study reveals that paedophile activity is nearly three times more prevalent in eDonkey than in KAD. While the study is limited by its focus on single-keyword queries and a static keyword list from 2010, it provides a robust methodology for cross-platform comparison. Future work should adapt these methods to modern encrypted messaging and streaming platforms to track the evolution of these behaviors.

Summary updated

Abstract

Peer-to-peer (P2P) systems are widely used to exchange content over the Internet. Knowledge on paedophile activity in such networks remains limited while it has important social consequences. Moreover, though there are different P2P systems in use, previous academic works on this topic focused on one system at a time and their results are not directly comparable. We design a methodology for comparing \kad and \edonkey, two P2P systems among the most prominent ones and with different anonymity levels. We monitor two \edonkey servers and the \kad network during several days and record hundreds of thousands of keyword-based queries. We detect paedophile-related queries with a previously validated tool and we propose, for the first time, a large-scale comparison of paedophile activity in two different P2P systems. We conclude that there are significantly fewer paedophile queries in \kad than in \edonkey (approximately 0.09% \vs 0.25%).
BibTeX
@misc{fournier2012comparing,
  author = {Raphaël Fournier and Thibault Cholez and Matthieu Latapy and Clémence Magnien and Isabelle Chrisment and Ivan Daniloff and Olivier Festor},
  title = {{Comparing paedophile activity in different P2P systems}},
  year = {2012},
  eprint = {1206.4167},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/1206.4167},
}
Figure