Re: (FWD) Results of stress testing the current version of the Onion Router

[Roger Dingledine <arma@mit.edu>]

On Thu, Sep 05, 2002 at 09:57:24AM -0400, Bruce Montrose wrote:
> That fixed the problem Roger. Performance under stress is still pretty bad, 
> but at least it doesn't crash now when I beat it up :)

Good to hear. :) Ok..it looks like we're getting around 3 successful
transactions a second, when it's under really heavy load, based on
those graphs?

Can you play around with it and see if you can find the highest steady
load that it can handle? What other questions can you answer, and what
other questions do we want to ask?

The limitation we're seeing is (I think) how quickly the first onion
router can build an onion. We plan to switch from 3DES to AES down the
road (once we get around to using the new openssl), but I don't think
that will help things much.

(A question for the applied crypto people here: is RSA slow compared
to other public key systems? Should we look at switching? What horrible
side effects will there be?)

--Roger