int-serv Cspecs (was Re: Last Call for draft-ietf-rolc-apr-00.txt )

[Curtis Villamizar <[e-mail]>]

In message <[e-mail]>fi>, Juha Heinanen writes:
> so there is a clear requirement to int-serv/rsvp that the user must be
> able to, in addition to qos, also specify at which cost he/she is
> willing to ask for such a qos.  is this already included?  in fact, the
> user should have capability to specify alternative sets of requirements
> and the network would pick the one it can support.
> 
> -- juha

Tspec, Rspec, and Cspec?  Where C is cost?  ;-)

I'm just a lurker on rsvp and int-serv.  I've yet to see a means of
communication cost requirements proposed.

There are clearly ommissions in the int-serv work.  If they ever
finish debating token-bucket and means or expressing and guarenteeing
delays, they'll probably move on to real issues.  :-)

You are welcome to propose a means to integrate cost into the Rspec or
porposed that rsvp adapt an additional Cspec.

Curtis

ps- maybe this thread should drop off rolc and continue on int-serv or
not at all.