Msg#: 3279042 posted 1:56 pm on Mar 13, 2007 (gmt 0)
You should talk to SkyRider, a startup that's helping advertisers get in front of P2P search queries. They have technology that allows them to 'convince' the peers on P2P networks to rank search results (including their sponsored ads) the way SkyRider wants.
Sequoia (#2-3 VC here in Silicon Valley) invested in them, and the team are folks who in the past started Narus and worked on NSA wiretapping projects - they know their networking protocols. CEO is Ed Kozel who sits on Yahoo's board.
I imagine that if they wanted to they could tell you the top 100K queries on the P2P newtorks.
Msg#: 3279042 posted 6:17 pm on Mar 15, 2007 (gmt 0)
I did something like otem is talking about a few years ago. I downloaded an open source client for that network that ran on Linux (gnutella, I think). I turned up the bandwidth settings to pass lots of search messages. Then I redirected the search monitor to a text file. Easy. Megabytes of search data were soon on my drive. Then I used grep and other utilities to extract and count the occurences of terms I was interested in.
The new gnutella is gtk-gnutella - it has a graphic interface. I used the text-based client, as it had easier access to the underlying API.