Comments on: TorrentFreak on “Solutions to the BitTorrent Problem” http://techliberation.com/2008/02/04/torrentfreak-on-solutions-to-the-bittorrent-problem/ Keeping politicians' hands off the Net & everything else related to technology Tue, 14 Feb 2012 19:26:00 +0000 hourly 1 http://wordpress.org/?v=3.2.1 By: BitTorrent Dir http://techliberation.com/2008/02/04/torrentfreak-on-solutions-to-the-bittorrent-problem/comment-page-1/#comment-58395 BitTorrent Dir Tue, 24 Feb 2009 18:43:10 +0000 http://techliberation.com/2008/02/04/torrentfreak-on-solutions-to-the-bittorrent-problem/#comment-58395 <p>well, i think that nowadays is kind of stupid of thinking about a solution but what can we say..</p> well, i think that nowadays is kind of stupid of thinking about a solution but what can we say..

]]> By: enigma_foundry http://techliberation.com/2008/02/04/torrentfreak-on-solutions-to-the-bittorrent-problem/comment-page-1/#comment-40541 enigma_foundry Sun, 10 Feb 2008 20:46:35 +0000 http://techliberation.com/2008/02/04/torrentfreak-on-solutions-to-the-bittorrent-problem/#comment-40541 <p><i>Interesting. Number 4 (Limit the total connections allowed at one time per user) is exactly that Comcast appears to be doing.</i></p> <p>Well, Richard you are right if and only if you can accept the following statement: if protocol=bittorrent then connexions allowed = 0</p> Interesting. Number 4 (Limit the total connections allowed at one time per user) is exactly that Comcast appears to be doing.

Well, Richard you are right if and only if you can accept the following statement: if protocol=bittorrent then connexions allowed = 0

]]> By: eee_eff http://techliberation.com/2008/02/04/torrentfreak-on-solutions-to-the-bittorrent-problem/comment-page-1/#comment-45485 eee_eff Sun, 10 Feb 2008 20:46:35 +0000 http://techliberation.com/2008/02/04/torrentfreak-on-solutions-to-the-bittorrent-problem/#comment-45485 <p><i>Interesting. Number 4 (Limit the total connections allowed at one time per user) is exactly that Comcast appears to be doing.</i><br><br>Well, Richard you are right if and only if you can accept the following statement: if protocol=bittorrent then connexions allowed = 0</p> Interesting. Number 4 (Limit the total connections allowed at one time per user) is exactly that Comcast appears to be doing.

Well, Richard you are right if and only if you can accept the following statement: if protocol=bittorrent then connexions allowed = 0

]]> By: Jeff Darcy http://techliberation.com/2008/02/04/torrentfreak-on-solutions-to-the-bittorrent-problem/comment-page-1/#comment-40540 Jeff Darcy Fri, 08 Feb 2008 01:29:10 +0000 http://techliberation.com/2008/02/04/torrentfreak-on-solutions-to-the-bittorrent-problem/#comment-40540 <p>Fair queuing doesn't really help, because it's generally implemented to work on flows (port pairs) rather than hosts. You could make it work on host pairs instead, modulo some NAT issues, or even on sender only instead of sender/receiver pair, but only by changing router software. The "problems" involved here exist despite already-widespread use of FQ, and I even remember discussing this very issue with Bram while BitTorrent was still in development. <br /><br /> One way that FQ might be deployed effectively, though, would be to use WFQ where the weights are assigned based on sender and/or receiver flow/connection counts. More connections, lower weights, and you lose (to an appropriate degree) if there's congestion. Sharing information about connection counts between routers might make such an approach even more effective without making it unfair. <br /><br /> BTW, if one were to read the actual bills and not just the opinion pieces related to net neutrality, one might reach the conclusion that any such scheme would be prohibited. That's why I think net neutrality <em>as currently defined in those bills</em> would do more harm than good.</p> Fair queuing doesn’t really help, because it’s generally implemented to work on flows (port pairs) rather than hosts. You could make it work on host pairs instead, modulo some NAT issues, or even on sender only instead of sender/receiver pair, but only by changing router software. The “problems” involved here exist despite already-widespread use of FQ, and I even remember discussing this very issue with Bram while BitTorrent was still in development.

One way that FQ might be deployed effectively, though, would be to use WFQ where the weights are assigned based on sender and/or receiver flow/connection counts. More connections, lower weights, and you lose (to an appropriate degree) if there’s congestion. Sharing information about connection counts between routers might make such an approach even more effective without making it unfair.

BTW, if one were to read the actual bills and not just the opinion pieces related to net neutrality, one might reach the conclusion that any such scheme would be prohibited. That’s why I think net neutrality as currently defined in those bills would do more harm than good.

]]> By: Jeff Darcy http://techliberation.com/2008/02/04/torrentfreak-on-solutions-to-the-bittorrent-problem/comment-page-1/#comment-45484 Jeff Darcy Fri, 08 Feb 2008 01:29:10 +0000 http://techliberation.com/2008/02/04/torrentfreak-on-solutions-to-the-bittorrent-problem/#comment-45484 <p>Fair queuing doesn't really help, because it's generally implemented to work on flows (port pairs) rather than hosts. You could make it work on host pairs instead, modulo some NAT issues, or even on sender only instead of sender/receiver pair, but only by changing router software. The "problems" involved here exist despite already-widespread use of FQ, and I even remember discussing this very issue with Bram while BitTorrent was still in development.<br><br></br><br><br>One way that FQ might be deployed effectively, though, would be to use WFQ where the weights are assigned based on sender and/or receiver flow/connection counts. More connections, lower weights, and you lose (to an appropriate degree) if there's congestion. Sharing information about connection counts between routers might make such an approach even more effective without making it unfair.<br></br><br></br><br><br>BTW, if one were to read the actual bills and not just the opinion pieces related to net neutrality, one might reach the conclusion that any such scheme would be prohibited. That's why I think net neutrality <em>as currently defined in those bills</em> would do more harm than good.<br></br></p> Fair queuing doesn’t really help, because it’s generally implemented to work on flows (port pairs) rather than hosts. You could make it work on host pairs instead, modulo some NAT issues, or even on sender only instead of sender/receiver pair, but only by changing router software. The “problems” involved here exist despite already-widespread use of FQ, and I even remember discussing this very issue with Bram while BitTorrent was still in development.



One way that FQ might be deployed effectively, though, would be to use WFQ where the weights are assigned based on sender and/or receiver flow/connection counts. More connections, lower weights, and you lose (to an appropriate degree) if there’s congestion. Sharing information about connection counts between routers might make such an approach even more effective without making it unfair.



BTW, if one were to read the actual bills and not just the opinion pieces related to net neutrality, one might reach the conclusion that any such scheme would be prohibited. That’s why I think net neutrality as currently defined in those bills would do more harm than good.

]]> By: Aaron Swartz http://techliberation.com/2008/02/04/torrentfreak-on-solutions-to-the-bittorrent-problem/comment-page-1/#comment-40539 Aaron Swartz Thu, 07 Feb 2008 05:03:16 +0000 http://techliberation.com/2008/02/04/torrentfreak-on-solutions-to-the-bittorrent-problem/#comment-40539 <p>How come nobody is listing fair queueing? It seems like the obvious solution here.</p> How come nobody is listing fair queueing? It seems like the obvious solution here.

]]> By: Aaron Swartz http://techliberation.com/2008/02/04/torrentfreak-on-solutions-to-the-bittorrent-problem/comment-page-1/#comment-45483 Aaron Swartz Thu, 07 Feb 2008 05:03:16 +0000 http://techliberation.com/2008/02/04/torrentfreak-on-solutions-to-the-bittorrent-problem/#comment-45483 <p>How come nobody is listing fair queueing? It seems like the obvious solution here.</p> How come nobody is listing fair queueing? It seems like the obvious solution here.

]]> By: Richard Bennett http://techliberation.com/2008/02/04/torrentfreak-on-solutions-to-the-bittorrent-problem/comment-page-1/#comment-40538 Richard Bennett Wed, 06 Feb 2008 21:34:29 +0000 http://techliberation.com/2008/02/04/torrentfreak-on-solutions-to-the-bittorrent-problem/#comment-40538 <p>Interesting. Number 4 (Limit the total connections allowed at one time per user) is exactly that Comcast appears to be doing.</p> Interesting. Number 4 (Limit the total connections allowed at one time per user) is exactly that Comcast appears to be doing.

]]> By: Richard Bennett http://techliberation.com/2008/02/04/torrentfreak-on-solutions-to-the-bittorrent-problem/comment-page-1/#comment-45482 Richard Bennett Wed, 06 Feb 2008 21:34:29 +0000 http://techliberation.com/2008/02/04/torrentfreak-on-solutions-to-the-bittorrent-problem/#comment-45482 <p>Interesting. Number 4 (Limit the total connections allowed at one time per user) is exactly that Comcast appears to be doing.</p> Interesting. Number 4 (Limit the total connections allowed at one time per user) is exactly that Comcast appears to be doing.

]]>