Re: Relative performance of http:// vs. svn://

>>>>> "Christopher" == Christopher Ness <chris@nesser.org> writes:

 Christopher> On Mon, 2005-09-19 at 09:57 -0400, Paul Koning wrote:
>> svnserve is consistently significantly faster. Specifically:
>>
>> checkout 20:37 vs. 32:52 switch 8:55 vs. 21:32 status 3:53 vs.
>> 5:21 update 6:30 vs. 9:01

 Christopher> Wouldn't you expect that? There is one less layer of
 Christopher> software overhead for your transactions to go through.

I didn't expect that. Neither side is CPU bound.

 Christopher> A big advantage of Apache is the Authz Access Control
 Christopher> List. But per-directory authorization is being worked
 Christopher> on for svnserve as well.

Based on earlier suggestions, I had turned off Authz entirely. That
doesn't make a difference in performance.

        paul

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Received on Mon Sep 19 18:01:25 2005