Dear authors, I’ve read the draft draft-dunbar-nvo3-nva-mapping-distribution-01.
In the draft, the NVE seems only passively accept the mapping information from NVA (push, pull or hybrid), and the NVA gets some mapping information from VM manager (sec. 7.2). That means NAMD works depending on some other protocol say NVA-VM manager protocol? Generally speaking, this same mapping information also exists in NVE (either configured or automatic created, or obtained via Hypervisor-NVE protocol). So NVE can distribute this mapping information to NVA directly by NVE-NVA protocol, than depending on the VM Manager-NVA interface, and the NVA then further distributes this mapping information to all other related NVEs in VN. And this is the main idea of the draft draft-gu-nvo3-vntp-00. (sorry for it was not properly prepared, it’ll be revised soon. ) Hope didn’t misunderstand. Best regards Zhongyu
_______________________________________________ nvo3 mailing list [email protected] https://www.ietf.org/mailman/listinfo/nvo3
