On Tue, May 22, 2007 at 09:30:11AM +0200, <BrokenClock>: ~> I've just quickly read this article, and thought you may consider it. It ~> describs a routing remplacement method that should greatly improve ~> networks efficiency. Here is the link: ~> http://science.slashdot.org/article.pl?sid=07/05/21/2227240&from=rss ~> <http://science.slashdot.org/article.pl?sid=07/05/21/2227240&from=rss> ~> Please let me know what you think about it. Should such distributing ~> method be used and usefull with something like netsukuku?
Thanks for the link! I really appreciate it. The network coding approach is clearly useful for multicast data flow, or any other case where there's a need of speeding up data transmission. It doesn't really concern the routing issues, although it is possible to implement a routing mechanism aware of coding strategies: http://arena.cse.sc.edu/papers/rocx.secon06.pdf Netsukuku uses the Caustic Routing, an optimized form of multipath routing, to distribute among several routes a same data flow: http://lab.dyne.org/Ntk_caustic_routing It's probably possible to integrate the network coding in the Causting Routing, this would give a better performance. Maybe it will even simplify its theory and implementation, because it won't be anymore necessary to privilege disjoint routes. To make the Netsukuku routing aware of the network coding it should suffices to "add" a coding metric. Right now there are the rtt, bwupload, bwdownload and average metrics. Also here, the search for disjoint routes won't be anymore so vital. However, an implementation of the network coding has to act at kernel level, because it needs to directly mangle the packets. We'll see. Best regards. -- :wq! "I don't know nothing" The One Who reached the Thinking Matter '.' [ Alpt --- Freaknet Medialab ] [ GPG Key ID 441CF0EE ] [ Key fingerprint = 8B02 26E8 831A 7BB9 81A9 5277 BFF8 037E 441C F0EE ] _______________________________________________ Netsukuku mailing list [email protected] http://lists.dyne.org/mailman/listinfo/netsukuku
