Greg Stein <gstein@lyra.org> writes:
> > I think we're
> > all out of silver bullets when it comes to checkout performance.
> > (Unless Tobias's latency claim turns out to be true.)
>
> Of course it is true. Always has been. We operate the HTTP pipe in
> half-duplex mode. That's been known for a while, and "HTTP pipelining" is
> the solution for that. Unfortunately, Neon doesn't have that feature :-(
>
> (pipelining [and using it in svn] was actually one of the desires and
> motivations for the ASF's "serf" project)
>
> I find it kind of funny that this is a "big discovery" :-)
>
> The WC performance still blows and should be optimized regardless of
> whatever network-related changes are made.
Yes, but if a network change can make the bigger difference -- Ben and
I are trying some experiments tomorrow -- then we should do that
first, and then see where we are.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Thu Oct 23 05:42:29 2003