Re: [Gzz] 29th, 2nd, 3rd, 4th and 5th (hh)

[hemppah]

Quoting "B. Fallenstein" <address@hidden>:

> address@hidden wrote:
> > Quoting "B. Fallenstein" <address@hidden>:
> > 
> > 
> >>Hi,
> >>
> >>address@hidden wrote:
> >>
> >>>-Started looking for information about "Gnutella++": hybrid system,
> where
> >>
> >>DHTs
> >>
> >>>and Gnutella-like approaches are merged. Or to be more specific: system,
> >>
> >>which
> >>
> >>>has efficient routing algorithm (comparable to DHTs), and in which users
> >>
> >>are
> >>
> >>>able to host resources locally (instead of mapping resources away from
> >>
> >>local
> >>
> >>>computer into the virtual address space)
> >>
> >>I don't understand this. What do you mean by 'mapping resources away 
> >>from local computer into the virtual address space', and how can it be 
> >>avoided?
> > 
> > 
> > This was what I meant:
> > 
> > There is a DHT based approach taken by e.g. Chord and CAN; in these
> approaches
> > values are stored at the nodes with identities "closest" to the key for
> the
> > (hash) value. Furthemore, if services are put on the computer that hosts
> the
> > corresponding key-value pair, the key-to-node mapping still means that
> this
> > computer is generally not owned by the user who decided to provide the
> > service(the value) to the network.
> 
> Hmmm. I think I don't understand well enough-- my initial reaction is, 
> "why bother? would that be worth it?", but I should really understand 
> first. What should I read?

For example, you don't want your data to be mapped to any other computer ?-)

I think http://iris.lcs.mit.edu/proposal.html is a worth of reading. Moreover,
you can read "About search methods, key revokation ..." thread which can be
found from p2p hackers mailing list:
http://zgp.org/pipermail/p2p-hackers/2002-December/thread.html


-Hermanni


-------------------------------------------------
This mail sent through IMP: http://horde.org/imp/