On Tuesday 28 October 2003 07:28 am, Some Guy wrote:
> --- Ian Clarke <[EMAIL PROTECTED]> wrote:
> > > You are suggesting that specailization makes no sense in small
> > > networks and that freenet does what is optimal. Ian, please think
> > > about this. In a tiny network were every node was connected to every
> > > other one, specialization still makes sense.
> > > Optimally each node would be specialized in a portion of hashspace
> > > proportional to its resources. Requests/inserts could work in HTL=1.
> >
> > No, optimally every node caches everything and there is no
> > specialization whatsoever.
>
> Well, I guess that's "optimal" for redundancy. It certainly isn't
> "optimal" from a storage or bandwidth point of view. Maybe that's the bush
> we're beating around.
>
> So when do you think it becomes "optimal" not to store everything on every
> node? Seems to me it couldn't be "optimal" to do this once N > HTL of
> inserts.
>
> This "optimal" word is a bit funny.
AFAICS, it becomes optimal when totalSizeOfDataInNetwork is significantly
greater than averageDatastoreSize. And even then, specialization is a matter
of degree.
--
"I love deadlines. I love the whooshing sound they make as they go by."
- Douglas Adams
Nick Tarleton - [EMAIL PROTECTED] - PGP key available
_______________________________________________
Devl mailing list
[EMAIL PROTECTED]
http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/devl
