Re: HTTP protocol v2: rethunk.

On Thu, Nov 6, 2008 at 9:18 AM, Ben Collins-Sussman
<sussman_at_red-bean.com> wrote:
> I'd have to defer this question to the serf experts. There's an
> unspoken assumption that saturating the pipe with parallel (and/or
> pipelined) requests is always a speed gain, but I actually don't know
> of evidence to back this up. Maybe it's out there, though.

For commits, the huge benefit will be to use pipelining - not multiple
connections. Currently, there is a CHECKOUT/PUT required for every
cycle - since the PUT can't begin until the CHECKOUT response is
parsed, this requires that the network connection is idle for large
amounts of time. This 'dead time' is bad - and is removed in the v2
design.

Multiple connections is likely not to be a *huge* win for commits. I
probably wouldn't even advocate doing that at first. -- justin

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe_at_subversion.tigris.org
For additional commands, e-mail: dev-help_at_subversion.tigris.org
Received on 2008-11-06 18:37:44 CET