gpg: Signature made Tue May 20 14:45:06 2008 UTC using DSA key ID E43DA450
gpg: Good signature from "Matthew John Toseland <toad@amphibian.dyndns.org>"
On Tuesday 20 May 2008 15:23, Michael Rogers wrote:
> Matthew Toseland wrote:
> > Hmmm, I thought you were arguing that the latency would be unacceptable
for a
> > message board system?
>
> I was arguing that you couldn't mix ten-second latency with ten-day
> latency in the same system. Usenet messages used to take several days to
> reach the furthest corners of the net because some connections were only
> active at night, and people still found it useful.
>
> > Also how would you prevent DoS?
>
> I'm not sure - hashcash might help, and maybe fair queueing to limit the
> scope of the attack to nodes near the attacker. A high-latency version
> of Freenet would also have to solve this problem.
True, but it would use different means to do it: some form of token passing.
>
> > Broadcast routing requires manual filtering, no? In order to prevent DoS?
>
> Usenet uses broadcast - I don't know how it deals with DoS but I guess
> most servers have a limit on the number of messages per user per day,
> and if a group gets really badly spammed people just curse and
> unsubscribe (so the network no longer needs to distribute the spam).
Hmmm, perhaps.
>
> > With passive requests, a message system would likely have almost exactly
the
> > same performance on a high latency Freenet as on a broadcast-routed
network.
>
> Only if you eliminated round-trips (eg redirects, splitfiles, SSK pubkey
> caching), which would require different data formats and a different
> protocol, meaning the high-latency and low-latency networks would have
> separate content and separate applications. At that point I think it
> would be fair to ask why the two networks were bundled together under
> the same name.
Depends on the usage, doesn't it? For widely subscribed message boards,
redirects wouldn't delay things more than one local round-trip, no?
IMHO the web (even if it's high latency) plus usenet is more useful than just
usenet.
However, high latency Freenet may not be feasible because of swapping...
>
> Cheers,
> Michael
