Re: Re: DHT usage and requirement lists

Am Mittwoch 06 August 2008 21:51:53 schrieb Michael Rogers:
> Good idea. Perhaps the TTL should be equal to the publisher's uptime (up
> to a maximum of 24 hours)? That means the republishing overhead will
> decrease exponentially the longer the publisher is online.
>
> We should probably consider the overhead caused by very short-lived
> peers - perhaps peers shouldn't start publishing their files in the DHT
> until they've been online for 15 minutes (then republish after 15, 30,
> 60, 120, 240 minutes as suggested above).

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

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

I read in some paper (don't know anymore where) that in Kad a TTL of 12h would 
be a significant improvement over 24h, with publishing redundancy halved from 
the current value. So we might want to think twice before setting the max TTL 
to 24h. Also please remember that in non-US countries dynamic IP addresses 
are the norm and often change every 24 hours. If a node is started at a 
random time, it's average first uptime will be 12h - and many people shut 
down their computers during the night. 

How big is the average uptime of Gnutella clients currently? 

Best wishes, 
Arne

-- My stuff: http://draketo.de - stories, songs, poems, programs and stuff :)
-- Infinite Hands: http://infinite-hands.draketo.de - singing a part of the 
history of free software. 
-- Ein Würfel System: http://1w6.org - einfach saubere (Rollenspiel-) Regeln. 

-- PGP/GnuPG: http://draketo.de/inhalt/ich/pubkey.txt


[Non-text portions of this message have been removed]