Re: [Detnet] Éric Vyncke's No Objection on draft-ietf-detnet-mpls-over-ip-preof-10: (with COMMENT)

["Andrew G. Malis" <[e-mail]>]

Eric,

Regarding your comment:


## Is MPLS over UDP the only solution ?

While using MPLS over UDP is indeed a valid solution for sequencing the
packets, I wonder whether RFC 8939 could have been updated/extended to also
add
ordering, this would probably have less overhead. Was this discussed in the
WG?
If so, why not adding this justification in the draft ? (note I may have
missed
something obvious :-O )


Just my 2 cents as a co-author.

One of the goals of the WG was to reuse existing IETF technology as much as
possible. This is documented in RFC 8938:

3.1.1.  Data Plane Technology

   The DetNet data plane is provided by the DetNet service and
   forwarding sub-layers.  The DetNet service sub-layer generally
   provides its functions for the DetNet application flows by using or
   applying existing standardized headers and/or encapsulations.

Given that, it was very straightforward to take advantage of RFCs 7510 and
8964. While we perhaps could have saved a few bytes in the header by
rolling our own solution, reusing RFCs 7510 and 8964 simplifies both the
specification and implementation, and follows the design goal in RFC 8938.
Plus this technology isn't intended for environments (such as low-power or
very slow networks) where a few bytes in the header would make much of a
difference.

Perhaps adding a reference to that section of 8938 would satisfy your
comment.

Cheers,
Andy


>