I can't make sense of this.  Using DCCP has nothing to do with strong security
so I presume that dbh is really referring to flow control and saying because
flow control may be an issue, everyone MUST implement DCCP (which means that a
whole load of RFC are invalid because they require UDP and do not require
DCCP!).

I see this as issue 8, for which I see the resolution as DCCP is better than UDP
and SHOULD be used when it is available (which I do not expect it to be).  I do
not think that we should go beyond that.  Where for example is MUST implement of
DTLS/DCCP in the latest SNMP transport document,
draft-ietf-isms-dtls-tm with which dbh has been heavily involved?

Tom Petch

----- Original Message -----
From: "Chris Lonvick" <[email protected]>
To: <[email protected]>
Sent: Monday, June 07, 2010 5:21 PM
Subject: [Syslog] Issue 13 - DCCP?


> Issue 13 - DCCP?
>
> Tom Petch wrote:
> vvv
> Another issue that came up from the IESG is the relative roles of UDP and
> DCCP as a substrate.  In this context, the discussions on tsvwg which Lars
> is steering about SCTP, DCCP and UDP make interesting reading, with some
> contributors asserting that the only way to get a packet through a complex
> network is with UDP, that SCTP and DCCP are (comparative) failures that
> just don't get recognised widely enough.
>
> Certainly my (limited) view is that UDP is the MUST HAVE, the one that
> will give maximum interoperability so while DCCP is technically superior,
> making it the MUST implement will simply cause this I-D to be ignored by
> most.
>
> I haven't seen any response from Lars on this issue.
> ^^^
>
> DBH responded:
> vvv
> Lars provided advice quite a while back. I concur with his advice.
>
> Implementers MUST implement support for DCCP (which should require
> minimal changes from support for UDP),
> so that if DCCP is available, and the operator chooses to use DCCPP,
> the implementation will work with DCCP.
>
> I view this as very similar to our standard security posture - stroing
> security is MUST implement, so it is available if the operator wants
> it. The operator is not required to use it.
> ^^^
>
> ACTION:  None - I think this is resolved.
>
> _______________________________________________
> Syslog mailing list
> [email protected]
> https://www.ietf.org/mailman/listinfo/syslog

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

Reply via email to