The -01 revision of draft-ietf-idr-tunnel-encaps has the following
changes from the -00 revision:
- By popular request, it has been written in such a way as to obsolete
RFC5512. This means that anything useful in RFC5512 had to be
incorporated into the new draft. I would welcome opinions on whether
this was done correctly.
- Two new sub-TLVs are specified: "MPLS Label Stack" and "Prefix-SID".
I would welcome opinions on whether these are useful or not. (I'm
pretty sure that the first is useful, the second is more speculative.)
- If you are familiar with deployed uses of the Encapsulation Extended
Community, the Color Extended Community, or the Router's MAC Extended
Community, it may be worth checking section 4 to make sure that the
draft does not introduce any problems.
- I wish more folks would take a critical look at section 8, which is
primarily about the use of VXLAN/NVGRE/VXLAN-GPE together with labeled
address families.
I would also be interested in hearing if anyone has an opinion on the
utility of using this sort of mechanism to signal IPsec tunnels. Once
RFC 5512 is obsoleted, RFC 5566 ("BGP IPsec Tunnel Encapsulation
Attribute") will need to be revised. It might be possible to generalize
that in such a way as to facilitate the secure interconnection of two
private ASes across the public Internet. Comments on whether RFC 5566
takes a reasonable approach would be welcome.
_______________________________________________
BESS mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/bess
