P2P searching trends: 2002–2004

This paper presents the trends of searching queries by users from peer-to-peer (P2P) networks over an 18-month period from July 2002 to January 2004. Four data sets of search queries collected from Gnutella were studied to describe the searching trends. Major findings include (1) the percentage of duplicate queries ranging from 34% to 68% of total queries; (2) an increase in non-English queries; (3) an approximately half of searching queries specified for video or audio file types; (4) the stop word “the” accounting for one-third of total stop words; (5) the shift of queries from audio to video; and (6) P2P users demanding for timely entertainment and porn materials. Based on the findings, it is worthwhile for P2P developers to consider (1) system design that allows effective searching using multiple languages; and (2) techniques that eliminate stop words for faster searching.