On Feb 2, 2012, at 8:47 PM, Jesse Gross wrote: > On Tue, Jan 31, 2012 at 10:34 AM, kmestery <kmest...@cisco.com> wrote: >> I'm curious what the current state of the VXLAN patches for Open vSwitch is. >> I saw an email from Justin on 1-24-2012 indicating the patches are being >> reworked at this time. I've gone ahead and forward ported the patches from >> last fall which Ben posted to the list, I've put them on github here: >> >> https://github.com/mestery/ovs-vxlan >> >> Before I spend much more time on this, I wanted to see if others are working >> on these so as not to duplicate effort. Is anyone currently working on these >> patches? If not, I will proceed to test them out and hope to submit them for >> inclusion into Open vSwitch. > > The reason why the vxlan patches went out and then never included is > because they're not quite sufficient to implement the full protocol > including multicast learning. Initially it was hoped that > implementing the wire format would be a good first step but in reality > some more infrastructure work needs to be done first so we put it on > hold while we figured it out. > > The goal of the infrastructure work would be to provide information > about the tunnel header in the flow lookup and then enable the > destination for outgoing packets to be set more flexibly from > userspace. It would then be possible to implement the current port > based mechanism, vxlan learning, and other things all from userspace > using this mechanism. > > Is this something that you'd be interested in taking a shot at? If > so, I can provide some more details about what I was thinking.
Yes, I can take a shot at this. Any details on what you are thinking here would be great, feel free to send to the list or unicast to me. Kyle _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev
