RE: Another question on draft-ietf-tewg-diff-te-proto-01.txt

[Francois Le Faucheur <flefauch@cisco.com>]

Dave,

At 13:24 08/07/2002 -0400, David Allan wrote:

> You mean to support the
> engineering of individual class types in an E-LSP? An E-LSP by definition
> carries mulitple class types....

Not quite.
By definition and E-LSP is CAPABLE of carrying traffic from multiple
Traffic Aggregate. But it is perfectly fine to transport traffic from
differnet Traffic Aggregates on differnet E-LSPs (i.e. one Traffic
Aggregate per E-LSP).

When you do not do DS-TE, you woudl generally transport multiple Traffic
Aggregates over E-LSPs.

When you do DS-TE you would either:

        

- use
L-LSPs

        

- use
E-LSPs with traffic from a single Traffic Aggregate

        

- use
E-LSPs with traffic from multiple Traffic Aggregates BUT you then have a
few constraints (see REQTS draft) such as not being able to perform
separate Constraint Based Routing of each TA on  a given 
E-LSP.


Cheers

Francois

> cheers 
> Dave 
>
> > -----Original Message----- 
> > From: Nabil Seddigh
> [mailto:nseddigh@tropicnetworks.com]
> > Sent: Monday, July 08, 2002 12:29 PM 
> > To: Allan, David [CAR:NS00:EXCH] 
> > Cc: te-wg@ops.ietf.org 
> > Subject: Re: Another question on
> draft-ietf-tewg-diff-te-proto-01.txt 
> > 
> > 
> > 
> > Dave, 
> > 
> > Mechanisms to support multiple class types on an E-LSP
> were proposed 
> > in this WG. Some providers and vendors expressed
> interest in 
> > such schemes. However, in Utah, the WG felt that
> standardizing 
> > multiple OA-ELSP would be premature until the basic
> DS-TE 
> > mechanisms are standardized and shown to be viably
> deployable. 
> > 
> > Updated versions of the related drafts can be found
> below: 
> > 
> >
> http://www.ietf.org/internet-drafts/draft-ganti-tewg-diffserv-
> > multicos-elspreq-00.txt 
> > http://www.ietf.org/internet-drafts/draft-ganti-mpls-diffserv- 
> > elsp-02.txt 
> > 
> > Best, 
> > Nabil Seddigh 
> > 
> > 
> > 
> > > This implies a 1:1 relationship between TTs and LSPs, this wouldn't 
> > > necessarily be true for an E-LSP. Now the mechanisms don't 
> > currently exist 
> > > to support associating multiple class types/bandwidth 
> > constraints for an 
> > > E-LSP, but if traffic trunk is exclusively an L-LSP 
> > concept, couldn't we pin 
> > > this down a bit more explicitly..? 
> > > 
> > > thanks 
> > > Dave 
> >