Everyone, knowing the performance of other anonymizing services such as Tor or I2P, probably thought while reading the headline something like: What the heck? Well truely this seems like an inconsistence, but yesterday I tested a JonDos (formerly known as AN.ON) pay cascade with relation to speed. Surely it wasn’t any cascade. As you can see, I’m the middle-mix operator of the tested cascade what could lead to promotion reproaches. Anyhow we’ll see what comes back but whats certain: I was surprised myself and there’s the shot (click here for fullsize)…
The JAP client version used was 00.10.004-dev, which was released a few days ago. Probably the Firefox precaching boosted the kb/s up a bit, but the main influence had for sure the location of the downloaded file which was a tagesschau.de VOD.
Here’s the traceroute from mix to filesource:
traceroute to 126.96.36.199 (188.8.131.52), 30 hops max, 52 byte packets
1 qwe.x.x.x.de (213.X.X.X) 0.331 ms 0.198 ms 0.187 ms
2 yxc.x.x.x.de (213.X.X.X) 0.474 ms 0.184 ms 0.179 ms
3 asd.x.x.x.de (213.X.X.X) 24.686 ms 0.726 ms 0.546 ms
4 xe-2-1-0-271.fra20.ip.tiscali.net (184.108.40.206) 3.623 ms 3.539 ms 3.773 ms
5 ge-6-0-0.fra10.ip.tiscali.net (220.127.116.11) 3.677 ms ge-5-1-0.fra10.ip.tiscali.net (18.104.22.168) 3.605 ms ge-7-0-0.fra10.ip.tiscali.net (22.214.171.124) 3.727 ms
6 126.96.36.199 (188.8.131.52) 3.644 ms 3.844 ms 3.655 ms
It’s for sure not the average performance and admitted, definitely not applicable to website retrieval. Nowadays website complexity multiplies latency times. They of course can and are beeing cached, but an anonymized site view will never be as fast as a normal one.