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

["Dimitry Haskin" <dhaskin@axiowave.com>]

Francois,

> Dimitry,
> 
> Thanks very much for your input. It's a good way to kick start the
> discussion on BC models. My thoughts below:
> 
> At 13:38 12/07/2002 -0400, Dimitry Haskin wrote:
> >Francois,
> >
> >Consider this posting as my input into the informal discussion that I
> will
> >no be able to attend.
> >
> >The Russian Doll BC model is a  fine model in its own rights. However
it
> may
> >not be suitable as the default BC model that is mandated for
diff-serv
> >capable MPLS devices.  My concern is that, due the nested nature of
its
> >bandwidth constraints, the RD model does not make much sense if
> preemption
> >of established LSPs is not used as a part of a given diff-serv
> deployment.
> 
> 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).

I think you would agree that the bouncing off LSPs is necessary for the
RD model to work. For simplicity consider two CTs, CT1 and CT0, and two
BCs, BC1 and BC0 where BC1 < BC0. According to the model, the sum of
normalized CT1 and CT0 reservations cannot exceed BC0 and at the same
time CT0 can be booked up to BC0. So when accepting a new CT1 call would
cause the combined CT1 and CT0 reservation to exceed BC0, the only way
to comply with the model is to preempt some of the CT0 LSPs.

> 
> 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 respectfully disagree. I believe that (i) and (ii) can be achieved
simultaneously with (iii). Preemption of LSPs at a lower service level
in order to allow LSP calls at a higher service level is not the only
way to achieve (i) and (ii). An alternative would be to allow a
potential service degradation of the already established LSPs at a lower
service level without totally removing (often in an arbitrary way) any
particular LSP from the service. Given my example, the sum of normalized
CT1 and CT0 reservations would be allowed to exceed BC0. One may even
argue that in the face of bursty traffic this might be a good way to
achieve higher link utilization when hard service guaranties at a lower
service level are not required.

> I believe SPs have requirements for (i) and (ii) and don't have a
problem
> with using preemption, which is an existing TE mechanism.
> I believe the Russian Doll model is a very good way to achieve (i) and
> (ii).  Yes it involves preemption to do so, but that is exactly why
> preemption was specified for (ie let some LSPs which arrive later
bounce
> earlier LSPs).
> More generally, as discussed in draft-lefaucheur-tewg-russian-dolls-
> 00.txt,
> the Russian Dolls model appears as a very efficient way to address the
> most
> important SP requirements (using preemption where necessary).

Except that the RD model also may impose use of the preemption when it
is not necessary or even desirable. Would you agree?

Regards,
 Dimitry

> 
> I don't think it would make any sense to sacrifice (i) or (ii) just
for
> the
> sake of not using preemption. Do you?
> 
> Cheers
> 
> Francois
> 
> 
> >The question really boils down to whether support of preemption is to
be
> >required (at least in the IETF framework) from the diff-serv capable
MPLS
> >devices. So far support of preemption was optional. Moreover, IMHO,
> >diff-serv can be deployed in quite satisfactory way without use of
> >preemption. Furthermore, preemption of established LSPs may not be
even
> >desirable in some deployments.
> >
> >Given that the RD BC model implicitly requires use of preemption for
it
> to
> >work, the posted question should be explicitly addressed one way or
> another
> >before the RD model can be selected as the model that every diff-serv
> >capable LR must implement.
> >
> >Thanks,
> >   Dimitry
> >
> > > -----Original Message-----
> > > From: Francois Le Faucheur [mailto:flefauch@cisco.com]
> > > Sent: Friday, July 12, 2002 5:59 PM
> > > To: te-wg@ops.ietf.org
> > > Subject: IETF54- Informal discussion on BC Model for DS-TE
> > >
> > >
> > > Hello,
> > >
> > > TEWG will not meet in Yokohama/IETF54.
> > >
> > > I'd like to invite the people who will be in Yokohama and
> > > have thoughts on
> > > default Bandwidth Constraints model for DS-TE, to get
> > > together for a very
> > > informal discussion on that topic.
> > >
> > > Let's meet in the lobby of the main hotel at 17:30pm on Monday.
> > >
> > > See you there
> > >
> > > Francois
> > >
> > >