[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

Re: New passive performance metrics in Tor




Hi Everyone, 

it is been a while since we discussed this topic here: 
As Björn already pointed out in an earlier message, we are interested in the fraction of bidirectional traffic in tor. 

Therefore we developed a metric with Karsten that tries to capture the necessary information. 
He implemented this metric into Tor (thanks for that): https://gitweb.torproject.org/karsten/tor.git/shortlog/refs/heads/bidistats2

After checking the code, we can confirm that it looks very good and matches exactly our intentions.   

In order to get wide range results, we would prefer to merge the code into the main Tor distribution. 
However should this constitute any trouble, we would be satisfied with an implementation on only a few selected running relays too. 


If you have further questions, please do not hesitate to ask. 

Cheers, 

Florian and Björn. 



Thanks, Björn!  That was quite interesting.

It's not urgent, but when you can, could you please have a look at the
measurements implemented in
https://trac.torproject.org/projects/tor/ticket/1819 and confirm
(there or here is fine) whether the way Karsten's code is measuring
"bidirectionality"  is indeed the way that would be helpful in your
work?

in fact, Florian and me discussed that with Karsten before (not on this
list, though) - the implemented metric resulted from this discussion. I
am therefore quite confident that this is actually what we need.

Nevertheless we will take a look ASAP.


-- 
Florian Tschorsch
Mobile and Decentralized Networks
Heinrich-Heine-University
Universitätsstr. 1, D-40225 Düsseldorf

Building 25.12, Room 02.43
Phone +49 211 81 11635
Fax +49 211 81 11638