[nvo3] Open issues for draft-ietf-nvo3-bfd-geneve

xiao.min2@zte.com.cn Mon, 30 November 2020 07:09 UTC

Date: Mon, 30 Nov 2020 15:09:32 +0800
Message-ID: <202011301509324745848@zte.com.cn>
Mime-Version: 1.0
From: xiao.min2@zte.com.cn
To: nvo3@ietf.org
Content-Type: multipart/mixed; boundary="=====_001_next====="
Archived-At: <https://mailarchive.ietf.org/arch/msg/nvo3/Qx7jBqM0B0iqMraqlDhd9fP5gHc>
Subject: [nvo3] Open issues for draft-ietf-nvo3-bfd-geneve
Precedence: list

Dear NVO3 WG,






This mail intends to share some open issues for draft-ietf-nvo3-bfd-geneve, and ask for your review and comments.



For BFD-over-Ethernet-over-Geneve encap in Figure 1, if the VAP of the terminating NVE has no IP address, for IPv6, whether the range ::ffff:127.0.0.0/104 or the address ::1/128, should be used? Within draft-ietf-bfd-vxlan-16 which is in RFC-Editor's Queue, the range ::ffff:127.0.0.0/104 is used, nevertheless we also heard opinion that the address ::1/128 is the only loopback address for IPv6.


For BFD-over-Ethernet-over-Geneve encap in Figure 1, if the VAP of the originating NVE has no IP address, whether the IP address of the originating NVE, or the address 0.0.0.0 for IPv4 and the address 0:0:0:0:0:0:0:0 for IPv6, should be used? In my view, considering the possible address overlap between the tenant space and the NVE space, the latter choice seems more secure.


BFD Demand Mode and BFD Echo function, whether there is a requirement to support one or both of them? Within draft-ietf-bfd-vxlan-16 which is in RFC-Editor's Queue, only BFD asynchronous mode is in the scope, both BFD demand mode and BFD echo function are outside the scope.


BFD Discriminator exchange mechanism, whether this draft needs to define a mechanism for BFD discriminator exchange? In my view, this mechanism is optional, and it can be achieved by EVPN, Geneve-Ping, Configuration, etc.






Best Regards,


Xiao Min

[nvo3] Open issues for draft-ietf-nvo3-bfd-geneve xiao.min2