<div>>>A can speak to B, B can speak to C, but A can't speak to C<BR></div> <div>this is why you have a relay/super node...to be the go between between A nad C; this is also why creating p2p systems is very difficult...b/c you'll have to test this along with tons of other different senarios.</div> <div><BR><BR><B><I>Michael J Freedman <mfreed@cs.nyu.edu></I></B> wrote:</div> <BLOCKQUOTE class=replbq style="PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #1010ff 2px solid">On Fri, 10 Mar 2006, Jeff Rose wrote:<BR><BR>> Date: Fri, 10 Mar 2006 11:08:57 +0100<BR>> From: Jeff Rose <ROSEJN@GMAIL.COM><BR>> Reply-To: Peer-to-peer development. <P2P-HACKERS@ZGP.ORG><BR>> To: Peer-to-peer development. <P2P-HACKERS@ZGP.ORG><BR>> Subject: Re: [p2p-hackers] clustering<BR>> <BR>> It seems like people are always putting arbitrary restrictions on p2p systems <BR>> and simulations in terms of connectivity, but is this really necessary? <BR>> Unless you are
trying to use NATed nodes (assume we can punch or route <BR>> through a neighbor),just about any pair of computers on the internet can be <BR>> neighbors. In essence the internet is a fully connected overlay graph.<BR><BR>The problem is that "just about every" and "every" node being able to <BR>communicate are not quite the same thing. Indeed, it's precisely the <BR>difference in these two assumptions which actually raises a lot of <BR>problems when actually deploying DHTs in the wide-area.<BR><BR>We recently presented a short paper at WORLDS '05 which discusses the <BR>real-world problems that arises from non-transitivity in Internet routing:<BR><BR>A can speak to B, B can speak to C, but A can't speak to C<BR><BR>as we all independently discovered from running CoralCDN, OpenDHT, and i3. <BR>(Firewalls and NATs are actually an easier problem that this, as they <BR>express routing constraints much more
symmetrically.)<BR><BR>http://www.scs.stanford.edu/mfreed/docs/ntr-worlds05.pdf<BR><BR>I sent an email about this paper to this mailing list a few months ago, <BR>and my apologies for the repeat. However, as our main audience for this <BR>paper was actually meant to be the hacker community, as opposed to the <BR>academic one, I thought it bears re-mention.<BR><BR>--mike<BR><BR>-----<BR>www.michaelfreedman.org www.coralcdn.org<BR>_______________________________________________<BR>p2p-hackers mailing list<BR>p2p-hackers@zgp.org<BR>http://zgp.org/mailman/listinfo/p2p-hackers<BR>_______________________________________________<BR>Here is a web page listing P2P Conferences:<BR>http://www.neurogrid.net/twiki/bin/view/Main/PeerToPeerConferences<BR></BLOCKQUOTE><BR><BR><BR>You don't get no juice unless you squeeze<br>Lemon Obrien, the Third.