[p2p-hackers] Keep Alive and network performance

[SIMON Gwendal RD-MAPS-ISS]

In Solipsis, we also use a unique message for both "heartbeat" (confirm aliveness) and "keepalive" (maintain NAT hole-punching). A french saying is "One stone, two hits !" :-)

We also intended to adjust heartbeat periodicity to the "link value". Some links are essential to the network connectivity, so they deserve a small periodicity. Some links are valuable because of some criteria (in our case, the peer is graphically displayed). A "NAT hole-punching" periodicity is enough. Finally, some links are optional. We just keep them alive with a long periodicity because (i) why not, (ii) handshaking may be costly.

This may be applied in other p2p networks : adjust keepalive periodicity to the number of shared files, to the redudancy (in DHT), to the estimated half-life of peer...


-- Gwendal
http://solipsis.netofpeers.net







> -----Message d'origine-----
> De : p2p-hackers-bounces at zgp.org 
> [mailto:p2p-hackers-bounces at zgp.org] De la part de Greg Bildson
> Envoyé : jeudi 9 février 2006 19:04
> À : Peer-to-peer development.
> Objet : RE: [p2p-hackers] Keep Alive and network performance
> 
> Rather than using a dedicated message for keepalives, 
> LimeWire first looks
> for message traffic of any kind on a connection.  If there is 
> traffic on the
> connection during a time window, then the connection is 
> assumed to still be
> alive.    If there has been no natural traffic, we will send 
> a keepalive
> ping to it from the hub/ultrapeer.  If there is still no 
> network traffic
> coming from the host after a recheck time, we assume the 
> connection is dead
> and kill it on the ultrapeer.
> 
> We were concerned both about unclean disconnects and 
> connections where the
> bandwidth was getting so choked that it was effectively dead.
> 
> Thanks
> -greg
> 
> > -----Original Message-----
> > From: p2p-hackers-bounces at zgp.org 
> [mailto:p2p-hackers-bounces at zgp.org]On
> > Behalf Of Matthew Gertner
> > Sent: Wednesday, February 08, 2006 3:04 PM
> > To: Peer-to-peer development.
> > Subject: [p2p-hackers] Keep Alive and network performance
> >
> >
> > Hi,
> >
> > We have a straightforward P2P topology with each peer in the network
> > attached to hub. We want to detect efficiently that a peer has
> > disconnected from the hub. The obvious way to do this would 
> be to use
> > our current TCP connection with a relatively short Keep 
> Alive value (a
> > few seconds so we could see the peer's presence updated 
> quickly if they
> > don't disconnect cleanly). Does anyone have a feeling for the
> > performance implications of this approach if many peers are 
> attached to
> > a single hub? Is there another approach that is recommend 
> (such as using
> > a custom UDP-based keep alive protocol)?
> >
> > Regards,
> > Matt
> > _______________________________________________
> > p2p-hackers mailing list
> > p2p-hackers at zgp.org
> > http://zgp.org/mailman/listinfo/p2p-hackers
> > _______________________________________________
> > Here is a web page listing P2P Conferences:
> > http://www.neurogrid.net/twiki/bin/view/Main/PeerToPeerConferences
> 
> 
> _______________________________________________
> p2p-hackers mailing list
> p2p-hackers at zgp.org
> http://zgp.org/mailman/listinfo/p2p-hackers
> _______________________________________________
> Here is a web page listing P2P Conferences:
> http://www.neurogrid.net/twiki/bin/view/Main/PeerToPeerConferences
>