Re: IETF54- Informal discussion on BC Model for DS-TE

[Nabil Seddigh <nseddigh@tropicnetworks.com>]

Francois,

> I don't follow you. Russian Dolls model does not "mandate" the use of
> preemption. It just uses it for what it has been specified for (ie bounce
> off LSPs when needed).
> 
> My impression is that it is just not possible to simultanesouly :
>          -(i) ensure bandwidth sharing (ie no bandwidth wastage)
>          -(ii) ensure bandwidth isolation (ie a CT cannot have some of its
> bandwidth taken by another CT)
>          -(iii) refuse to use preemption
> I don't think it would make any sense to sacrifice (i) or (ii) just for the
> sake of not using preemption. Do you?
> 

I agree with what Dimitry wrote and remains along the
line of what I posted about a month ago:

http://psg.com/lists/te-wg/te-wg.2002/msg00099.html

I think the primary tradeoff between the RD and MA models is in
(i) and (iii). (ii) is supported by both RD and MA.

The MA model is actually easier to conceptualize and understand
in relation to network operations - albeit somewhat wasteful
as it does not allow bandwidth sharing across classes. 
As such it would appear to be a better choice if a default model
is chosen. I would think the simplicity of operation would be 
a key criteria for a default model.

Operators who wish to deploy the RD model can of course do so
via configuring something other than the default model.

Nabil

Nabil