On Tue, 19 Nov 1996, Craig Bishop wrote:
> Martin Ibert wrote:
> >
> > Huh? I thought the whole discussion was about avoiding duplicate cache
> > objects in the two caches. The trade-off should therefore be between
> > inter-cache traffic and cache disk space. Out-link bandwidth shouldn't
> > be an issue.
> >
> > (Or did I get something wrong?)
>
> The issue really is getting the maximum number of unique objects in
> multiple tightly coupled caches. The current cache relationships do
> not allow this.
>
> As an aside.
>
> Bandwidth costs us megabucks, disk and processing power is cheap by
> comparison. Think US prices times at least 10 sometimes 20 (no kidding).
>
> Cheers, Craig
In that case wouldnt you want both caches to hold objects, even if they are
duplicated on both caches? To save those $$$ when in fact one of the caches
does go down.
?
-- The Fulcrum Consulting Group Peter Marelas - Consultant 12/10-16 Queen St, Melbourne VIC 3000,Australia Ph: +61-3-9621-2100 PGP Key -> finger maral@fusion.mel.sprint.com.au Fx: +61-3-9621-2724Received on Mon Nov 18 1996 - 06:24:19 MST
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:33:34 MST