Re: [Diffserv] Comments on the TCB of the Conceptual Model - msg 1 of 2

[Andrew Smith <[e-mail]>]

John,

The next version of the MIB (-04) will, indeed, reference the TCB concept. But
only for one specific function, that of determining which TCB a given classifier
element is located in (it is possible to have the same classification patterns
used multiple times along a given path through the functional datapath elements
- we had earlier overlooked this fact so we will need to add TCB as an index).

There are several fairly central concepts in the model draft - TCB is just one
of them (I'd put it slightly less central than the low-level functional element
taxonomy and the idea of hooking together elements with arrow-like plumbing). It
should be used by instantiations of the model (e.g. MIBs, PIBs, schemas) as they
see fit: the MIB happens to focus on a very low-level description of the
functional datapath elements so it needs to reference the TCB concept somewhat
less often (zero times in -03, one time in the coming-soon-to-a-list-near-you
-04). 

I can't speak accurately for the next PIB incarnation but I would imagine it has
a similar level of need for the TCB concept.

There's nothing wrong with this.

Andrew


John Strassner wrote:
> 
> Thanks for this Brian, but I also didn't see any defense of
> the TCB as is. If you want to leave the TCB in, then that's
> fine, but I think that its existence should be further
> validated. For example, I note that both the PIB and the MIB
> don't use the TCB.
> 
> Therefore, I don't see how the abstract can say that the
> model is a basis for building either if such a central
> concept of the model is not implemented in either the MIB or
> the PIB.
> 
> Plus, I would really like to see a response to my questions
> about the purpose of the TCB.
> 
> thanks in advance and regards,
> John