Re: svn commit: r1495419 - in /subversion/trunk/subversion/libsvn_ra_serf: options.c ra_serf.h serf.c util.c

On Wed, Jul 10, 2013 at 6:57 PM, Ben Reser <ben_at_reser.org> wrote:
>...
> I have about 160ms of latency to this server. Mostly because I'm on
> the west coast of the US and it's in the UK someplace based on it's
> domain name.

Human perception is right around the 250ms mark. User interface
designers use that metric for response time.

[ Google shoots for some huge percentile of searches to come in under
that. Once you cross the 250ms mark with *no response*, then the user
perceives a delay. In Google's case, that delay means a user goes
elsewhere, and Google loses money. They've demonstrated a specific
correlation between response time and revenue. It's a bit frightening
when you file an SEC annual report that says "our revenue is tied to
our HTTP response time". (okay, they didn't file that, but they
*could*) ]

Of course, you can play certain tricks, like:

$ svn ls $URL
<... 150ms ...>
Listing of $URL:
<... 200ms ...>
$FILES

In any case. In your example, with a 160ms latency, then RTT*2 is past
that barrier. The user will notice a lag.

That said: there may already be other lag, so RTT*2 might be 320ms of
a 5s response time.

>...

Thanks for running these numbers. I can see that it might also be
constructive to add some kind of connection/request profiling into
ra_serf that we could use to direct optimizations in the future.

Cheers,
-g
Received on 2013-07-11 06:58:16 CEST