[core] On the technical approach for CoAP message aggregation

[Carles Gomez Montenegro <[e-mail]>]

Dear CoRE WG,

In the CoRE WG session at IETF 122 (Bangkok), we received some comments
regarding the technical approach to support message aggregation in
draft-gomez-core-coap-bp-03, currently based on the Payload-length option.
An excerpt of the relevant part of the minutes can be found at the end of
this message.

The main point is trying to avoid the need to parse possibly several
options in each individual message (that belongs to an Aggregate message)
to determine the payload length of each such message.

To this end, there might be different approaches, including making the
Payload length option number to be zero or another low number (as mentioned
briefly in IETF 122), as well as modifying the actual CoAP header format.

One consideration is that, while message aggregation is currently defined
in the CoAP over BP draft (and we can define a specific CoAP header format
for CoAP over BP), perhaps message aggregation could be used over other
transports (e.g., UDP) as well.

We were wondering: would there be any knee-jerk reaction or any preferred
technical approach (e.g., among the ones mentioned) to support CoAP message
aggregation?

Thanks in advance!

Cheers,

Carles and Anna


------------- Excerpt from the CoRE WG session at IETF 122 ------------

CA: I like message aggregation. If it works out like that technically,
then aggregation with the Payload Length option might be an instant
candidate for -core-corr-clar.
CG: Yes, this may work outside this Internet Draft.
CB (on chat): +1 (Even though I'm still not sure why this isn't part of
the header...)
CB: I mentioned previously that, for CoAP over TCP, we changed the CoAP
header. Putting a Payload Length option is not so different, but it
requires parsing all the options in the individual messages in order to
parse the aggregate message.
CA (chat only): That was also my knee-jerk reaction; that is what I
meant with "not sure on technical implementation".