> If I understand correctly, work is now ongoing to create a separate
> implementation of HNCP?

It's just a private hack, but might end up being useful.  The initial plan
was to spend a week-end explaining DNCP to myself, but the week-end has
now become four days.

I'm doing full DNCP right now (with keepalives, without unicast endpoints,
and omitting one micro-optimisation that I happen to disagree with), and
just enough HNCP to get hnetd to speak to me.  1500 LOC, not counting MD5.

> This would be a good step to address my concern I have voiced privately
> to the authors that not enough people have gone through the document.

Yes, it would have been a good thing if some large private-sector actor
with an interest in home networking had sponsored an independent
reimplementation before the DNCP WGLC.

> I personally think this WGLC is premature.

I'm actually more concerned by DNCP.  HNCP is extensible -- if we get it
wrong, we can deprecate some TLVs and add new ones without breaking
compatibility.

Not so DNCP -- DNCP requires precise state synchronisation, so any trivial
change is likely to require a whole new protocol.  Plus there's no version
number in DNCP, so the only way to evolve it is to pick a new port
(picking a new multicast group will not be enough in case clients want to
participate in both variants).

I've got a large list of DNCP issues, which I'll probably be able to write
up today or tomorrow.

-- Juliusz

_______________________________________________
homenet mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/homenet

Reply via email to