Re: Review of draft-ietf-trill-over-ip-08

[Joe Touch <[e-mail]>]

Some observations:

- the title is misleading; this is TRILL over UDP, not trill over IP.

- the use of two different ports invites some potentially unintended
problems, e.g., selective blocking of the control vs. data plane. IMO,
given that TRILL's purpose is to extend Ethernet (not IP), this service
would be better served using a single port and differentiated
encapsulated traffic by whatever method TRILL nodes use internally.
Otherwise, this spec needs to include specific description of unexpected
behavior, e.g., data frames on the IS-IS port and IS-IS frames on the
data port.

- regardless of whether one or two ports are requested, this doc should
provide the needed information for IANA (e.g., a service name and
description compliant with RFC6335).

- the section on MTU handling might benefit from informationally citing
intarea-tunnels, and consider using the recommendations there. In
particular, it's not sufficient to assume IPv4 supports 576 byte MTUs
(that's the minimum receiver reassembly MTU, not the transit MTU). That
section should also address issues of PMTUD and PLMTUD.

FWIW.

Joe