Re: [squid-users] Squid usage

robin.garner@iname.com wrote:

> Intuitively it seems that you could make use of a multiprocessor box by creative use of a cache hierarchy ?

You can.

> This assumes of course that the CPU load of a proxy-only squid
> is significantly lower than one that manages a large cache.

My measurements seems to indicate this is the case, but I have yet to do
fully conclusive measurements.. (the box I have is overpowered in CPU
wrt disks..)

> Also, I would expect that you could improve overall response times
> on a loaded system like this by _decreasing_ the size of the
> cache_dirs. I'm assuming that cache management overhead increases
> (logarithmically ?) with the size of the cache.

Actually the management overhead stays approximately the same, but there
is a huge difference in memory usage.

> Given usage patterns that I've seen at a few sites, the hit rate
> between a 10GB cache and a 20GB cache is not really significant
> - not compared with the 0.6 second cache hit time - I generally
> see times in the 0.05-0.1s range (and not even on a real CPU).

Depends on the amount of users you have and the bandwidth..

> In general, I think the key statistic we should look at is
> the speedup, ie (all_request_time -cache_miss_time)/all_request_time.
> In this context, attacking the time per cache hit is going to bring
> better results than a few extra 0.1% on the cache hit ratio.

Well.. you also need to compare the miss latency times with what you get
without a proxy, or else the whole measurement is moot as you do not
have a reference.

--
Henrik Nordstrom
Squid Hacker