Clip2 just completed a study of the OpenNap network -- a network of index servers and peers that work similarly to those operated and distributed by Napster, Inc., but the system supports exchange of multiple file types, not just audio. Numerous unaffiliated parties operate OpenNap index servers advertised for public access by Napigator.
The number of OpenNap servers fell by over 75% in the period from late February to late April 2001. The number of simultaneous users among these servers declined from 100,000 in late March to less than 50,000 by late April, and the number of files available similarly dropped about 50% from its peak in March. Finally, the average file size among files shared by OpenNap users rose by over 20% between early February and late April.
The number of OpenNap servers advertised and available for public access has declined considerably since late February. The Wall Street Journal and CNET reported on February 23 that the Recording Industry Association of America had sent letters of complaint to ISPs providing service to OpenNap server operators, and the decline may be a result.
Because the user population is not spread evenly among servers, it has suffered less decline by proportion than the server population, although the dropoff remains considerable.
The typical number of files per user, around 225, has remained relatively stable. The file population has roughly varied with the user population as a result.
|What do these statistics mean? Is the RIAA's letter-writing campaign having a chilling effect?|
|Post your opinion|
To gather data for this study, Clip2 retrieved an OpenNap server list from Napigator multiple times per day. Our OpenNap monitor then made independent connections to the listed OpenNap servers and requested the numbers of connected users and available files. OpenNap servers may be linked in such a way that they report numbers that reflect multiple views of a single population. To correct for this duplication, we assumed that servers operated on the same domain were linked, and we averaged their populations instead of summing them. For each metric, we averaged over the multiple values observed each day and thus obtained representative daily values.
The coming and going of servers on the Napigator list introduces substantial variations into our data, and we therefore restrict our conclusions to trends on the timescale of 1-3 months. Our measurement of the server population is smaller than Napigator's because we exclude servers on the napster.com domain and those which repeatedly fail to accept a connection. Our study does not address the population of OpenNap servers not advertised at Napigator.
Kelly Truelove is an independent research analyst who, via Truelove Research, covers peer-to-peer technology with a focus on P2P content search, storage, and distribution networks. He is regarded as a leading expert on consumer file-sharing systems, which he covers with a data-driven approach.
Copyright © 2009 O'Reilly Media, Inc.