Re: [core] Fwd: New Version Notification for draft-gomez-core-coap-space-00.txt

[Christian Amsüss <[e-mail]>]

Hello Carles,
hello CoRE,
(deepspace@ removed b/c this is more CoAP technical)

On Wed, Dec 20, 2023 at 07:54:59AM +0100, Carles Gomez Montenegro wrote:
> no proposal has been made to add support of Forward Error Correction
> (FEC) to CoAP. 

There has been some discussion in hallway meetings and in the RIOT
community chats on FEC for block-wise transfer (which have indeed not
resulted in any written down proposals). This would not only be
applicable to deep space, but also to terrestrial multicast channels
when used for firmware updates.

Discussion there was mainly focused on regular and not Q-Blocks, because
the latter have a very tight set of constraints for their applicability,
and because mechanisms introduced there (eg. the
application/missing-blocks+cbor-seq type) can just as well be applied to
regular block-wise transfer.

One aspect that was discussed there was the choice of transmission
sequence and sharding, where memory and code complexity requirements are
traded against robustness of the protocol. Another trade-off to be
considered is fast forwarding: For the applications that have been
discussed previously, relaying through a proxy was highly relevant, and
that proxy can not only terminate retransmissions, but also participate
in FEC. If the proxy can not spool the whole resource, a wider spread of
FEC (which is desirable for robustness) leads to the proxy being unable
to participate in FEC. (As a complete space layperson, I imagine that
such a proxy would be highly capable though, and could easily do all FEC
recovery -- images of space stations receiving transmissions and
forwarding them to ground stations once they pass over them comes to
mind).

BR
Christian

-- 
To use raw power is to make yourself infinitely vulnerable to greater powers.
  -- Bene Gesserit axiom