Hi, I always wanted to share our results here, but I always forgot:). I guess, then, now it's time for that.
We have recently built a community-based framework (called NFPA) with one very aim in our mind: to investigate the pure performance of software-based (v)NFs in different environments and use-cases. After a measurement is conducted, it is possible to share the results with the community to make them to be ``officially published''. The main database has recently been purged, and only filled up with decent reproducible results, hence the amount of data is not as huge as we want at the moment, but it seems to be a good start. For instance, a stable OVS 2.5.0 with DPDK 2.2.0, running on 1 core of an Intel Core i7-6700 x86-based server, trying to switch 40G traffic based on 100 and 100.000 destination MAC addresses, respetively (i.e., acting an l2-switch) with an Intel XL710 NIC achieves the following results: goo.gl/za8iVV For more information about NFPA, the framework and the applications beneath, usage information, built-in use-cases and traffic traces, and for everything else, please visit: http://nfpa.tmit.bme.hu/ NFPA is open-source, and accordingly it's completely free. So, please use it, share results and do feedbacks! Moreover, it is being maintained and developed continuously. Cheers, Levi On 02/09/2017 12:13 AM, Kumuthini Ratnasingham wrote: > > Hi there, > > > Wondering does anyone have any report on > > > standard OVS , DPDK performance numbers ? > > > Thanks, > > Kumuthini > > > > _______________________________________________ > discuss mailing list > [email protected] > https://mail.openvswitch.org/mailman/listinfo/ovs-discuss
_______________________________________________ discuss mailing list [email protected] https://mail.openvswitch.org/mailman/listinfo/ovs-discuss
