Hi David, Thanks for clarifying. All my queries were in relation to problem statement "identifying and resolving non-unique/duplicate NVE identifiers", as observed in few vendor deployments. I think this will require end to end support between NVA<->NVE and NVA<->NVA with respect to any (extensions to existing or new) control protocol.
Regards, Saumya. From: <Black>, David <[email protected]<mailto:[email protected]>> Date: Monday, March 7, 2016 at 9:05 AM To: sadikshi <[email protected]<mailto:[email protected]>>, "[email protected]<mailto:[email protected]>" <[email protected]<mailto:[email protected]>> Cc: "[email protected]<mailto:[email protected]>" <[email protected]<mailto:[email protected]>>, "Black, David" <[email protected]<mailto:[email protected]>> Subject: RE: Mail regarding draft-ietf-nvo3-arch Hello Saumya, > This one is specific to deal with problem of "non-unique/duplicate NVE > identifiers" configured either due to erroneous configuration OR To err is human - and that's one of a myriad of possible configuration mistakes ... listing all of them seems pointless. > OR In case of two Distinct NVO3 data-center networks connected over a DCI/WAN > gateway to each other as mentioned in section "9. Federated NVAs" That's actually covered in Section 9: "IP-G would then decapsulate received traffic from one NV Domain, translate it appropriately for the other domain and re-encapsulate the packet for delivery." > * Some more elaboration required in section "9.1. Inter-NVA Peering" > on what kind of filters/mappings for ensuring this uniqueness As section 9.1 says: " To support peering between different NVAs, an inter-NVA protocol is needed. The inter-NVA protocol defines what information is exchanged between NVAs." If you have such a protocol design or design framework, please "send draft" to the WG, as how to design such a protocol is outside the current scope of the architecture draft (IMHO). Thanks, --David From: Saumya Dikshit (sadikshi) [mailto:[email protected]] Sent: Sunday, February 14, 2016 2:07 PM To: [email protected]<mailto:[email protected]> Cc: [email protected]<mailto:[email protected]> Subject: Mail regarding draft-ietf-nvo3-arch Hello Authors, I have a query if the following needs a specific mention as a requirement or as a potential problem in the nvo3 architecture document: This one is specific to deal with problem of "non-unique/duplicate NVE identifiers" configured either due to erroneous configuration OR In case of two Distinct NVO3 data-center networks connected over a DCI/WAN gateway to each other as mentioned in section "9. Federated NVAs" * Is there a requirement where-in this problem can to be explicitly mapped to in any of the architecture requirements in the document. * Some more elaboration required in section "9.1. Inter-NVA Peering" on what kind of filters/mappings for ensuring this uniqueness * This one is specific as the next hop (Overlay tunnel end-point) related information is duplicated. Thanks Saumya.
_______________________________________________ nvo3 mailing list [email protected] https://www.ietf.org/mailman/listinfo/nvo3
