Re: help over "rsync vs cvsupd perfmance"

["Justin C. Sherrill" <justin@xxxxxxxxxxxxxxxxxx>]

On Mon, December 25, 2006 10:12 am, Saverio Iacovelli wrote:
> I find, in the DragonFly project page, the following
> matter:
>
> - Benchmark rsync vs. cvsupd for getting source code
> updates
>
> Ok, I would like to test benchmark, but I need help to
> choose software for benckmarking and help to formulate
> test cases.
>
> Someone has suggestions?

I think I was the one who came up with that idea.  Testing it would be
relatively simple.  Set up two machines, networked together.  Set up
cvsupd and rsyncd on one machine, both using the same files as a
repository.  (all of DragonFly src would bet a good example)  On the other
machine, bring all those files down with rsync, and time how long it
takes.  Delete the files, and do it again with cvsup.  Repeat both a few
times just in case something happens to change the numbers during a run.

You'd have to set up rsyncd and cvsupd, neither of which I've done, and
you'd need two systems that were otherwise not busy, on a local network.

You could benchmark against one of the public sites that offers DragonFly
src both through cvsup and rsync, but the variability of the Internet path
from you to them and back would mess up the scores, and the relatively
lower speeds may mask any differences in the two protocols.

Other