Re: [core] WG interest in Sleepy Node topic

[Thomas Fossati <[e-mail]>]

Hi Akbar,

On Fri, Nov 8, 2013 at 2:36 AM, Rahman, Akbar
<[e-mail]> wrote:
> Once again we ran out of time on the agenda to discuss the WG interest on
> the topic of Sleepy Nodes

yes, I'm sorry about that.

> However, Carsten did suggest that we carry the discussion that we would have
> had in the meeting forward to the WG list.  The key questions that we need
> WG feedback on are:
>
> 1.       Does the WG want to keep discussing Sleepy Nodes in CORE?  The
> Sleepy Node support could be in:
>
> a.       CoAP Protocol
>
> b.      CORE Link Format
>
> c.       CORE Resource Directory

IMHO if we want to safely stay within the CoRE scope, we should keep
sleepy/intermittent handling at the CoAP protocol and Link Format
level (with RD used as a backup/complement to base Link Format
discovery).  So, to me this should be a.+b. with optional c.
And this is completely possible -- as documented in the Publish I-D.

> 2.       Sub-Question:
>
> a.       Does the WG want to see some immediate development of the Mirror
> Server concept (which seems to have near unanimous support and is well
> correlated to Sleepy Node support)?

No, for two reasons:
1) before adopting any solution I think we should look at the
different proposals on the table (there are quite a few) and be able
to match them against real use cases so to decide what is the best fit
for the case (or even better, for a non-trivial subset of cases);
2) I think Mirror belongs to the layer just above CoAP and L-F, and so
possibly out of CoRE scope.

Cheers (and thanks again for keeping the discussion alive).