ietf
[Top] [All Lists]

Re: [PCN] Re: WG Review: Congestion and Pre-Congestion Notification(pcn)

2007-02-21 03:28:11
On 2007-02-20 15:31, Fred Baker wrote:

On Feb 20, 2007, at 8:15 AM, Georgios Karagiannis wrote:

I assume that you also have no objection on using the DSCP fields for
this purpose.

actually, I do, at least in some ways that they might be used. The AF service (RFC 2597) is specifically designed to do as you say; EF isn't. setting the DSCP on an EF packet in a way that nullifies its EF behavior to communicate the onset of congestion mostly nullifies the EF service (which doesn't actually do much of anything when there is no congestion to perturb packet timing). So I would want any use of the DSCP to be coordinated with the other services that the datagram hopes for.

Exactly. The allowed uses of the DSCP are constrained by RFC 2474
and PCN can't go outside those constraints without causing unpredictable
breakage. That needs to be explicit in the charter, I believe.

While I'm writing, consider this assumption:

(A) these components are deployed in a single DiffServ domain,
where all boundary and interior nodes are PCN-enabled and
mutually trust each other

That is fine from the point of view of the diffserv architecture,
which is strictly domain-based. However, it isn't a get out of
jail card for security: BCP 61 (RFC 3365) still applies. The
comment about mutual trust should probably be removed.

    Brian

_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/ietf