Re: [Spud] related draft in TSVWG

On Mon, Jun 27, 2016 at 2:14 PM, Joe Touch <touch@isi.edu> wrote:
> Hi, all,
>
> The following draft may be of potential utility in SPUD/PLUS:
>
>      draft-touch-tsvwg-udp-options
>
> It describes an extension to UDP to allow for trailing options at the
> transport layer.
>

Hi Joe,

>From the draft:

"This length field is typically redundant with IP datagram length and
header length information as follows:

                  UDPlen = IPlen - IPhdrlen - 8

As a result of this redundancy, the UDP length field can be used in other ways."

"typically" is a key word here; this is not a requirement of UDP. As
long as the UDP length is less than or equal to IP payload length it
would be a valid UDP packet. The draft exploits this characteristic,
however that does not preclude that packets with smaller UDP length
than IP payload length already exist (maybe even using the same trick
to extend UDP!). Seems like that leads to a potential ambiguity?

Tom

> NOTE: I'm not sure what space makes the most sense; for on-path info,
> IMO IP is a more appropriate location than transport, but IF there is
> info that is useful to exchange at the transport layer in the header,
> this might be useful for UDP.
>
> It will be briefly presented by the chairs in TSVWG.
>
> FYI.
>
> Joe
>
> _______________________________________________
> Spud mailing list
> Spud@ietf.org
> https://www.ietf.org/mailman/listinfo/spud