On Mon, May 05, 2003 at 06:47:21PM -0400, Edward J. Huff wrote: > Working out whether or not a design is vulnerable to various attacks is > complicated... When hashed file storage is isolated from Freenet into a > separate network of nodes ostensibly devoted to research into hash > collisions (call it SETHC, the Search for Extra Terrestrial Hash > Collisions), we need some way to be able to translate a CHK into one or > another pair of SETHC hash codes. But yet this translation _must_ > remain secret, or else someone can request deletion of the hash codes. > I think there must be a way to do that, using an intermediate server > which forwards the request to SETHC in an anonymous way.
Umm, no. Read up on CHKs. We hash the file to get the encryption key, then we encrypt the file again, then we hash the result to get the routing key. The only reason it is readable is that we keep the encryption key, and publish it in the URL. If we do not publish the URL, for example if we are a node and not a client, it is random data for all practical purposes. > > -- Ed Huff _______________________________________________ devl mailing list devl at freenetproject.org http://hawk.freenetproject.org:8080/cgi-bin/mailman/listinfo/devl
