On Mon, Nov 10, 2014 at 09:22:03AM +0100, Martin Vizvary wrote: > @Ben - Now, i do understand the implementation and why there is such > short passive timeout. Are there any performance tests about the impact > of number of records in datapath flow table?
Of course--we don't just do this kind of thing to upset NetFlow users. In fact, you are the only person who has ever mentioned that this is a problem for NetFlow. Why can't (don't) the collectors aggregate records? I don't think there are published performance results related to tuning these parameters. We do tons of measurements internally when we're tuning OVS algorithms, but it's a big step from running some tests to writing a paper describing them. You can probably find some information in the OVS commit log, though, about the benefits of individual changes. _______________________________________________ discuss mailing list discuss@openvswitch.org http://openvswitch.org/mailman/listinfo/discuss
