Benson, I am a little bit confused about tag for VAP concept. In my understanding, tag for VAP is equivalent to tag for VN since VAP is (logical) port per VN thing according to arch draft?
If we assign a tag per VM, I agree that duplication by hairpin problem in indirect case does not exist any more. However hairpin-ed multicast packets has to be unicast with different Q-tag to the intented VMs? That would be tedious. Another approach could be having two tags. One for VN, another for tNVE. Hairpin-ed packets has to be delivered to certain tNVE first and then to VMs in certain VN attached to that specific tNVE. Yizhou -----Original Message----- From: Benson Schliesser [mailto:[email protected]] Sent: Monday, March 23, 2015 3:17 AM To: Liyizhou Cc: Anoop Ghanwani; [email protected]; [email protected] Subject: Re: [nvo3] VDP GroupID vs. VNID in draft-ietf-nvo3-hpvr2nve-cp-req-02 Hi, Yizhou. Thanks for your note. Liyizhou wrote: > I think we need to figure out how external NVE perform the hairpin > forwarding in this case before deciding if any other extension is > required. Personally I prefer not to enable hairpin forwarding in any > indirect connection case. Otherwise we still need some intelligence on > intermediate classic bridge B1. Is this true if we use a distinct Q tag for each VM? I.e. if we use the tag to identify a VAP, and do *not* think of this as Q-tag = VN, then we should be able to switch through intermediate devices and "hairpin" forward on the nNVE when appropriate. Is this correct and/or am I missing something? Cheers, -Benson _______________________________________________ nvo3 mailing list [email protected] https://www.ietf.org/mailman/listinfo/nvo3
