Re: Re: DHT usage and requirement lists

Arne Babenhauserheide wrote:
> How high is the cost of publishing? This protocol would publish quite often, 
> where long-lived nodes would neet to publish only once. 

Good point, we could save bandwidth by starting at a higher value if the
publisher knew it was likely to be online for a long time.

> Doesn't each Gnutella node already know its default (daily) uptime? At least 
> Phex nodes do. 

Do Phex peers keep any record of how their uptime varies, or do they
just keep a running average? I'd be interested to know whether
individual peers exhibit the same pattern as the network overall, ie
lots of very short sessions and a few very long sessions. If that turns
out to be the case then the length of the current session would be a
better predictor than the length of the average session.

Cheers,
Michael