Re: Subversion trunk (r1078338) HTTP(/WC?) performance problems?

On Thu, Mar 10, 2011 at 3:02 PM, Greg Stein <gstein_at_gmail.com> wrote:

> That was an explicit design choice. ra_serf should not be using the
> REPORT style of query. Those GETs are cachable by local proxy caches,
> reducing the need to (re)fetch the content over a WAN. The big REPORT
> response is never cachable.

I think the proxy cache is more theory than reality especially given
that any site that is SSL will not be cached by a proxy. That said, I
think the new pristine store offers us abilities in the future to not
download files we already have in the cache and the Serf style will
work much better with that concept (which of course is just another
cache).

> I'm also a little curious about serf using more CPU, as our testing a
> few years ago showing serf to be faster and with less CPU. Mebbe
> something changed since then :-/

I could be wrong, but I think all of the cases about the CPU are
related to WC-NG and not Serf. The high CPU is also seen with Neon.
This is more of a 1.7 vs 1.6 issue.

-- 
Thanks
Mark Phippard
http://markphip.blogspot.com/