On Jan 27, 2011, at 09:52, Lars Eggert wrote:
all new protocols should
be security-capable
Sure.
How is this relevant?
In some protocols, there is value to use them without communication security
(think TLS) for some applications, and with communication security for others.
We used to distinguish these two cases using two ports, now we use a single
port plus per-connection negotiation like STARTLS.
I think the draft is trying to encourage this conversion, and I agree with
this, at least where latency is less relevant.
With UDP-based protocols, it is harder to do this.
Please look at section 7.3 of
http://tools.ietf.org/html/draft-ietf-core-coap-04.html#section-7.3
and tell us whether this is how you would like this to be handled for UDP-based
protocols in the future.
If not, we may want to add to the guidance in the (tsvwg) draft.
Gruesse, Carsten
_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/ietf