Re: Kad User hash

From: Ben Grodsky <>
Date: Thu, 26 Apr 2007 18:29:14 -0700


How the data is collected, I'm assuming for purposes of the statician's understanding, should be discussed directly between Jay and the staitician. My understanding is there's a meeting early next week at your offices between y'all, Jay and Randy. Kad, as far as I understand it, doesn't show either supply or demand, but rather an echoed supply dependant on demand --- suppland/dupply. But I might be confusing kad with something from gnutella now. It's best that the scientists discuss it directly in person with one another.


----- Original Message -----
From: Skinner, Andrew (NBC Universal) <>
To: Ben Grodsky
Sent: Thu Apr 26 18:17:20 2007
Subject: RE: Kad User hash

Hey Ben,


Weird! I was just thinking about this and an email appeared.


I ran the sampling issue by Aaron and he said it’s okay as long as we see everything you see. There are some questions about how the data is collected though. Are user hashes collected from supply, demand, or both? Also, is it possible to get some sample sets of the ed2k hashes and kad hashes that are being collected? I’d like to do a comparison and a crosscheck to make sure that the same hashes are being collected on both sides.


For the time being go ahead and prioritize the eDonkey feed. Thanks for seeing this through.








From: Ben Grodsky []
Sent: Thursday, April 26, 2007 5:30 PM
To: Skinner, Andrew (NBC Universal)
Subject: Kad User hash



The kad user hashes are being recorded now and will be in a feed soon. In terms of priority, would you want first an eDonkey feed or a kad feed (kad is totally separate from eDonkey, though eMule allows users to connect to kad)? Again I'd like to reiterate the sampling issues you'll have in comparing bayTSP kad user hashes to MediaDefender ones will basically make it impossible to track users repeatedly infringing your copyrights.


Received on Fri Sep 14 2007 - 10:55:57 BST

This archive was generated by hypermail 2.2.0 : Sun Sep 16 2007 - 22:19:46 BST