RE: [Seamoby] CAR Discovery Requirements

Hi all,

Can anybody make clear to me why this separation of requirements is
really necessary? In the issues draft, we have identified three cases 
(anticipated, dynamic, and hybrid) to address the problem of CAR
discovery, 
and to my understanding the requirements cover all three. I would urge 
not only to suggest a separation but really to outline a proper
definition of
(separate) requirements so that everybody can see that this separation
makes sense. 

But just as a reminder (please look into the issue draft). This, i.e.,
CAR discovery, is merely about identifying GAARs including their
capabilities,
whereby the the spectrum and exact definition of the latter are not
within
the scope of the protocol, to enable either the AR (anticipated case) or
the MN (dynamic case) to determine a set of ARs, namely the CARs, that
meet 
MN's requirements.

HOW (and by what parties) to determine this set, e.g., by signalling the
MN's requirements to the AR for the anticipated case, and how to select
a 
TAR, i.e., the final candidate for handover, are out of scope of this 
protocol.

Hence, I don't understand how billing and accounting comes in here and
why anything should be considered separately for AR-AR or AR-MN. The
problems
to be addressed are similar for all three cases, namely anticipated,
dynamic, and hybrid. I admit that the decision as to how to select the
TAR 
would probably heavily depend on operator's policies. But this is
(again) not in
the scope of what we are discussing here. 

For instance, if you want to keep other ARs out of the decision of TAR
selection,
then implement your TAR selection algorithm appropriately. It does not 
at all influence the CAR discovery. It is rather a local (adminstrative)
implementation issue. And that's how it is defined in the issues
draft. 

In conclusion, John's comments are already considered by taking these
problems out of the scope of CAR discovery. How they still affect CAR 
discovery and its requirements, is unclear to me.

Regards,

Dirk

> -----Original Message-----
> From: ext James Kempf [mailto:kempf@docomolabs-usa.com]
> Sent: Monday, January 14, 2002 12:53 PM
> To: Hesham Soliman (ERA); John Schnizlein; Govind Krishnamurthi
> Cc: seamoby@ietf.org
> Subject: Re: [Seamoby] CAR Discovery Requirements
> 
> 
> Hesham,
> 
> 
> 
> > => This is exactly why I urge you (this wg) to clearly 
> > explain which protocol you are talking about. AR - AR or
> > AR MN ???
> > Clearly AR - MN lis an obvious way to discover all routers 
> > that the MN can hear. Inter/intra domain issues are 
> > not relevant. 
> > 
> > So please make it clear which part of the protocol
> > you're referring to. We should not always assume 
> > that we're discussing the AR - AR case. 
> > 
> You are correct, the protocol must apply to both cases,
> and the requirements should reflect that.
> 
> Perhaps it would make sense to divide the requirements
> document into three sections: a) common requirements,
> b) requirements for AR-AR case, c) requirements
> for AR-MN case.
> 
> 
>             jak
> 
> 
> _______________________________________________
> Seamoby mailing list
> Seamoby@ietf.org
> https://www1.ietf.org/mailman/listinfo/seamoby
> 

_______________________________________________
Seamoby mailing list
Seamoby@ietf.org
https://www1.ietf.org/mailman/listinfo/seamoby