On Thu, Feb 10, 2011 at 12:00:40PM +0100, Karsten Loesing wrote:

So, I implemented the new sanitizing algorithm that replaces bridge IPs with 10.x.x.x addresses derived from the original bridge IPs. I also sanitized the descriptors from November 2008 and performed an early analysis on the results.

The last remaining step is to sanitize our archives of bridge descriptors once again and make the sanitized data available. I don't expect to have new tarballs before next Tuesday.

The new tarballs are available here: https://metrics.torproject.org/data.html#bridgedesc

Before publishing the new tarballs, I'd like to invite people to review the new sanitizing process:

- A specification-like description of the sanitizing process is in Section 3 of https://metrics.torproject.org/papers/data-2011-02-10.pdf

- The early analysis results of bridges changing their IP addresses are here: https://trac.torproject.org/projects/tor/ticket/2435#comment:3

- The tarball with sanitized bridge descriptors from November 2008 is here (6.2M): http://freehaven.net/~karsten/volatile/bridge-descriptors-hashed-ips-2008-11...

Any questions, comments, and concerns are highly appreciated!

Feedback still much appreciated! Best, Karsten