Re: [dcp] Re: draft DCP charter for discussion

>so what is the relationship between DCP and the "congestion control 
>building blocks"/"congestion manager" approach that was all the rage a few 
>IETFs ago?

I think fractionally less than the relationship between TCP and CM.

You could implement TCP in such a way that it used CM.  Similarly you
could implement DCP in such a way that it used CM.  But you don't have
to in either case.

However, while TCP essentially has only one type of congestion control
feedback, DCP is more flexible - the feedback type and CC dynamics can
depend on the negotiated CCID.  Thus while you could use a CM to share
congestion state between TCP flows and/or DCP flows that use TCP-like
dynamics.  And you could use a CM to share congestion state between
DCP flows using TFRC feedback and dynamics.  But I wouldn't like to
speculate on how you'd share congestion state between a flow using
TCP-like feedback and dynamics, and one using TFRC feedback and
dynamics.

I don't think that the existence or non-existence of a CM greatly
affects the design of a protocol like DCP, although of course with
would impact the implementation.

Cheers,
	Mark

_______________________________________________
dcp mailing list
dcp@ietf.org
https://www1.ietf.org/mailman/listinfo/dcp