Post-Workshop project: Acquiring realistic user data

On the P2PIR workshop a lively discussion was centered about dynamics of a P2P network. Current research, unless peer joins and departures are ignored completely, relies on artificial churn models. The participants believed that a better evaluation requires realistic scenarios. The idea of a small project was born which should aim at creating such a test bed: Interested researchers from the P2PIR community should install a small program (yet to be created) which monitors the online times and queries posed, and make them available in an anonymous version. Queries to Google or CiteSeer (with a preference on the latter) should be logged by some kind of internal proxy, e.g. based on JAP. This issue will be discussed on the newly created post-p2pir2005 mailing liste.

You are cordially invited to joint the list and the discussion!