Re: How to optimize cacheing

G'day John!

John Saunders wrote:
>
> A last modified factor of 100% would means that if the object was say 2
> days old when it was first fetched into the cache, it would become stale
> after being in the cache for 2 days. A 50% factor would mean it would
> become stale after 1 day. Is this correct logic?

Close ... there are other tests before the LM_FACTOR one which may cause an
object to be marked FRESH/STALE. The logic is correct, but only applies if the
object makes it to the LM_FACTOR test.

> Having a minimum and maximum age seems pointless as when the object is
> in-between these two ages the algorithm thinks the object is stale.

The (AGE <= MIN_AGE) allows a cache admin to override any anti-cache header
settings that a site may have sent.

The (AGE > MAX_AGE) enforces an upper limit to the age of an object before an
IMS request is sent.

If an object has an expires time, which is greater than MIN_AGE, this is
honoured reguardless of MAX_AGE. I don't think you'll find many sites who issue
expires to set them for long periods of time. If they are, it's assumed that
the page doesn't change for that amount of time. LRU cleanup would purge these
obects if they started to clutter up space.

> Why have a max age when the object is already stale if it's older than the
> min age?

This only would occur if the object falied the LM_FACTOR test (and the MAX_AGE
test wasn't there).

> I don't seem to be able to understand the subtlety of the
> AGE > MAX_AGE check before the LM_FACTOR check. Can anybody clarify
> why this check is done and not just left to fall through to the end?

It stops objects whith a LM_FACTOR AGE > MAX_AGE being marked as fresh.

Later
Ed

-- 
Ed Knowles aka Jasper				   Phone : +61 2 9385 4962
E-mail: ed@fatboy.geog.unsw.edu.au	           Fax   : +61 2 9313 7878
            What I lack in morals I make up for in principles.