On Mon, Mar 10, 2014 at 06:49:55PM -0700, Ethan Jackson wrote: > Yep this is the same thing. I've successfully shown that this > algorithm is the best for our use case and was intending to implement > it myself. However, for many rather complicated reasons it became > important to have a prototype implementation quickly, so Jarno > graciously offered to pound out the implementation. > > On the issue of the algorithm choice. We tried a bunch of things, but > when we simulated it against some real firewall rule sets, we found > that the simplest is the best: a decision tree.
Can we have some kind of quantitative explanation of the benefits in the commit message? I know from our face-to-face discussions that you have measured how this reduces the number of megaflows required to cover the whole port space in various circumstances. _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev
