Thank you very much. I will test it when the download becomes available since I am now using squid.
I would like to see the new perf increase when I do test on Redhat 7.3 Kernel 2.4.19. Best regards Edward Millington BSc, Network+, I-Net+, CIW Associate Systems Administrator, Sr Cariaccess Communications Ltd. Palm Plaza Wildey St. Michael Barbados Phone: �1 246 430 7435 Mobile: 1 246 234 6278 Fax: � �1 246 431 0170 [EMAIL PROTECTED] www.cariaccess.com -----Original Message----- From: Igor Khasilev <[EMAIL PROTECTED]> To: [EMAIL PROTECTED] Date: Mon, 9 Sep 2002 14:28:44 +0300 (EET DST) Subject: [OOPS] Another version > Hi! > > Here is cvs from which you can take what currently is in another > version > > CVSROOT=:pserver:[EMAIL PROTECTED]:2401/usr/local/cvs > password: anoncvs > cvs co Plus > > > Briefly main differences(partially done): > > Thread pool in correct sense of the word. This will let programm to > serve > large number of connections using small number of threads (probably > will > help with linux 1:1 thread mapping). Maybe this will also give more > performance. > > C++ with all its pluses and minuses. > I never use this language actively. It was choosen for its pluses. > If somebody know it well - don't hesitate, any advise you will give > me will > be accepted with attention. > > Safe strings > > malloc - all documents memory is allocated from fixed pool. This will > give > more control over programm size in memory. > > Configurator, which can be used later for telnel-like or web-like > connects > and configuration. All options will be self-descripted, as I never > have time > to write separate files with descriptions. > > > It was succesfully compiled and run under Solaris. I didn't test it > under any > other system. > > I tested it with polygraph in simplest case: 1 client, 1 server (both > on the > same machine where proxy run), any caching disabled. Here is some > results: > > squid ~ 150-160 req/sec > oops ~ 220 req/sec > new code ~ 260 req/sec > > You shouldn't pay much attention to > these results, because new code don't contain any acl, groups, > modules and > so on processing. But I did it to check that I have code which at > start > works not slowly than any other proxy. > > The code in cvs is not usable in real work, it is as far from > production state as possible. > > Igor Khasilev | > PACO Links, igor at paco dot net | > > > > ===================================================================== > If you would like to unsubscribe from this list send message to > [EMAIL PROTECTED] with "unsubscribe oops-eng" in message body. > Archive is accessible on http://lists.paco.net/oops-eng/ ===================================================================== If you would like to unsubscribe from this list send message to [EMAIL PROTECTED] with "unsubscribe oops-eng" in message body. Archive is accessible on http://lists.paco.net/oops-eng/
