Another way to archive the web

Timo Burkard, Herodotus: A Peer-to-Peer Web Archival System, a master’s thesis submitted to MIT in 2002. [PDF] (Thanks to LIS News.) Distributed users donate unused CPU cycles and disk space to crawl the net and store its contents. Like LOCKSS, it uses duplicate copies to assure persistence. If every node could contribute 100 GB of storage, then (as of May 2002) Burkard estimated that it would take 20,000 nodes to archive the whole net. (PS: A quick Google search suggests that Burkard’s idea has been cited but not implemented. Does anyone know of an implementation? If it was tried and found wanting, does anyone know how it fell short?)
Posted by Peter Suber in the Open Access News