On Mon, 2013-09-16 at 21:59 +0100, Stephen Farrell wrote: > Interesting idea. I note you say that this might add 5-30 seconds > to the access time for a give site and more for a web page with > more than one TLS server cert needed. > > Is that something that could get quicker to the level where it'd > be acceptable do you think?
If more people ran Tor network exit nodes (using the existing Tor project software), and the existing nodes would be less of a bottleneck, it would probably get quicker. Also, it depends on the sphere configuration. In my initial experiments, one of the spheres was configured to accept only Tor exit nodes that are located in Africa or South America, but those countries have very few nodes currently (around 12 in total). For example, if you were located in Europe, and you configured all of the spheres to use european countries, only, it would probably be much quicker. > I suspect those kinds of latency would be a killer for browsers or > other applications using TLS. I agree that at this time the delay is a killer argument for enabling the feature by default for all users. However, there might be people who are very much worried about their security, who might be willing to accept such delays. Regards Kai _______________________________________________ therightkey mailing list therightkey@ietf.org https://www.ietf.org/mailman/listinfo/therightkey