I support the draft in this process, please see inline:

This begins a 2 week WG LC for draft-ietf-trill-transport-over-MPLS
(11/11/2017 to 11/25/2017).  Please consider if this draft is ready for
publication.   In this consideration please consider:

1)      Does TRILL need to run over MPLS?  Some data centers are
interconnected over MPLS.  Does this capability aid in deployment of TRILL?

[RP]: Yes to both questions.

2)      Is this specification ready for publication?

[RP]: I have no objection to it being published, but have some review
comments below. From an high-level/overview perspective, it seems
reasonably clear/complete in terms of providing a problem statement
and specifying the technologies needed to resolve the problem

3)      Do you know of any problems with this specification?


a . I think the draft may need to be more clear on how it maps an
incoming (control and/or data-plane) packet to a specific tenant in
either approach. Are pseudo-wires/MPLS circuits shared between
tenants? Specifically if two or more tenants have the same switch-id
on the Rbridge side, how does the solution differentiate between them
based on the incoming packet?

b. How does multi-topology Trill fit into this draft?

c. Typos in figures 2, 4. Under "Rbat2" on the left-side of each
figure, the label should probably be "Tenant 2 Site 1" (as opposed to
"Tenant 2 Site 2")?



Susan Hares

(co-chair, document shepherd)
trill mailing list

Reply via email to