Rob, Does OAM discussion happen in extensibility mailing list? If I understand your response correctly, OAM functionality will still be provided by Vendor switch/routers(at least for legacy switches) with hooks provided by controller. It might work well for BFD, however, I am not quite sure about Ethernet OAM? In addition for OF/hybrid switches will it be a similar approach or it would take a hybrid implementation in both switch and controller?
Thanks Ravi -----Original Message----- From: [email protected] [mailto:[email protected]] On Behalf Of Rob Sherwood Sent: Monday, August 22, 2011 11:20 AM To: Kingston Smiler Cc: [email protected] Subject: Re: [openflow-discuss] OpenFlow and OAM Mechanisms (BFD) On Mon, Aug 22, 2011 at 10:58 AM, Kingston Smiler <[email protected]> wrote: > Hi All, > The OpenFlow Spec 1.1.0 doesn't mention anything about the OAM mechanisms. > Please let me if anyone is working on the OAM aspects in OpenFlow. So, at least IMHO, OpenFlow is roughly orthogonal to most OAM mechanism. There do exist hooks from OpenFlow into a box's existing mechanisms, e.g., BFD, via the group table's "fast-failover" group, but OpenFlow proper focuses on forwarding. Independently, it's possible to do topology discovery strictly through OpenFlow's packet_in/packet_out messages, although there are scaling concerns about this approach and various people have talked about ways to better leverage the box's already existing OAM mechanisms, but they have never really taken off. Was there some specific use-case you were concerned about? I guess to more directly answer your question, there have been discussions about OAM but I'm not aware of any active proposals in the ONF. Strictly IMHO, but I haven't heard anyone discuss these issues as a major pain point. Hope that helps, - Rob . _______________________________________________ openflow-discuss mailing list [email protected] https://mailman.stanford.edu/mailman/listinfo/openflow-discuss _______________________________________________ openflow-discuss mailing list [email protected] https://mailman.stanford.edu/mailman/listinfo/openflow-discuss
