Measuring BitTorrent ecosystems

Abstract

Summary form only. BitTorrent is the most successful peer-to-peer application. In the last years the research community has studied the BitTorrent ecosystem by collecting data from real BitTorrent swarms using different measurement techniques. In this talk we present the first survey of these techniques that constitutes a first step in the design of future measurement techniques and tools for analyzing large scale systems. The techniques are classified into Macroscopic, Microscopic and Complementary. Macroscopic techniques allow to collect aggregated information of torrents and present a very high scalability being able to monitor up to hundreds of thousands of torrents in short periods of time. Rather, Microscopic techniques operate at the peer level and focus on understanding performance aspects such as the peers' download rates. They offer a higher granularity but do not scale as well as the Macroscopic techniques. Finally, Complementary techniques utilize recent extensions to the BitTorrent protocol in order to obtain both aggregated and peer level information. The talk also summarizes the main challenges faced by the research community to accurately measure the BitTorrent ecosystem such as accurately identifying peers or estimating peers' upload rates. Furthermore, we provide possible solutions to address the described challenges.