Re: [squid-users] Ideal cache placement (was Re: Why Squid is great (was: fourth cache off??))

Vivek starchily noted:
> It's also worth noting that our OEMs actually ship those
> configurations, whereas in the past, some of the Squid numbers
> have been for unstable configurations that would never have
> been placed in production - just benchmarketing.

You telling me that NONE of your OEMs have ever shaded anything?

> Plus, the iMimic throughput numbers are slightly less than 10x the
> Swell/Squid numbers, which, since I was talking about 10 Squids vs. 1
> iMimic, happens to model the relative load factors nicely.
>
> . . .
>
> From queuing theory, it's also very unlikely that
> those 10 caches will get a total performance that's 10 times a
> single cache.

Ahem. *Performance* has multiple components. Most of my assertions
are about latency, which I regard as a more important bottom-line
figure.

Even if you restrict your comment to throughput only, I don't think
it's true. If Internet traffic displayed smoothing upon aggregation,
you would be right. But it doesn't (fractal, etc etc.).

> > One point I am trying to make is that the central box is a box with
> > heavy load on it, and thus it has trouble keeping the latency down.
>
> That's contrary to what the benchmark shows:
> at 2700 requests/sec, the iCache 2500 showed hit times of 21.5ms and 2667ms,
> both of which are faster than the Squid entry running at 130 reqs/sec.

It is exactly what the benchmark shows. I was talking about the
iMimic iMimic box. Man, you seem so unwilling to talk about that box :->.

We could talk about the iCache 2500, seems like a nice box in many
respects, but it is more expensive than your own iMimic box. An
organization needs to have alot more users before they'll be looking
for an iCache 2500.

> It's true that for a particular instance, the hit/miss times will show
> some increase as the load is increased. However, the argument as I
> understood it was using a faster centralized cache versus slower
> distributed caches.

Well, yes, the *particular*instance* of a 10x load.

In the cacheoff, we have a snapshot of Squid at load <k>, and
the iMimic cache at load 10<k>. The cacheoff benchmark measured
*sustained* performance. Exactly what we're looking for.

-- 
Jon Kay        pushcache.com                      jkay@pushcache.com
http://www.pushcache.com/                             (512) 420-9025
Squid installation, maintenance, and coding       'push done right.'