On Mon, Dec 22, 2008 at 4:00 AM, PAPADIMITRIOU Dimitri
<[email protected]> wrote:
> I think we are in disagreement on the description of a cache:

You're right, we disagree. You held up a Golden Delicious apple, a
McIntosh apple and a Bartlett pear and said, "These are apples." The
pear isn't an apple. It may grow on a tree and it may have a similar
texture but a pear isn't shaped like an apple, it doesn't taste like
an apple and the bottom line is: it's not an apple.

Precomputing a FIB from a set of RIBs is not caching. It isn't even
memoization.

If you investigate starting with a faulty taxonomy, the "pear" will
serve as a false counter-example to many statements about apples which
are, in fact, true. You won't learn what you hope to learn, about
apples or caches.



>> Concrete example please. Is there an XYZ cache in IOS 12.9? Does JunOS
>> have a competing cache that works differently?
>
> There are past and ongoing experiments with such associative caches -
> does that make the case more real ? or is the reality limited to Cisco
> IOS or other commercial OS ?

I'll settle for a published research paper in which more than one type
of "associativity" is discussed in the context of caches excluding PC
memory caching.

I expect you to demonstrate that associativity is a caching issue in
more than one problem domain, hence an issue for caching in general
rather than an issue unique to PC memory caching. I'm not trying to be
dogmatic about this. You've claimed that associativity is a general
property of caches. I just want to see that claim supported.


>> If you replace 3.4 push/pull/hybrid with 3.4 pull local-central, pull
>> local-distributed, pull remote-central and pull remote-distributed,
>> you'll be closer.
>
> As said above pull local/remote from the forwarding point of view copes
> with the read operation. what about the operation needed in case of
> routing information change pushed downward onto the FIB (not triggered
> by a change in dest. addresses but routing topology) ?

I'm going to break off this branch of the conversation. We're starting
to delve into caching theory here. That's wonderfully interesting
stuff, I could go on about it all day, but it isn't on point. We're
supposed to be grounding the topic by discussing operational
experience with caches. For that, we should select a representative
sample of operational caching systems and then restrain the scope of
our discussion.

Regards,
Bill Herrin


-- 
William D. Herrin ................ [email protected]  [email protected]
3005 Crane Dr. ...................... Web: <http://bill.herrin.us/>
Falls Church, VA 22042-3004
_______________________________________________
GROW mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/grow

Reply via email to