RE: RE : [DNA] Review of draft-yegin-dna-l2-hints-01.txt

["Bound, Jim" <jim.bound@hp.com>]

I think it makes sense and good job on that one.  I don't think we can
define all theL2 detections though and have to leave it at a few
examples and make it clear the examples are conceptual.  It's the parts
we do with it at L3 and MIPv6 (IMO ok and don't care at all about MIPv4)
for detection that we need to make sure is baked and interoeprable and
no known bugs in operation from protocol perspective.  We might need an
API or two though once we "get-it" right and consensus exists.
My .87cents :--)
Thanks
/jim 

> -----Original Message-----
> From: Alper Yegin [mailto:alper.yegin@samsung.com] 
> Sent: Wednesday, July 14, 2004 6:59 PM
> To: greg.daley@eng.monash.edu.au; Bound, Jim
> Cc: 'NJEDJOU Eric FTRD/DMR/REN'; 'Bernard Aboba'; 
> dna@eng.monash.edu.au
> Subject: RE: RE : [DNA] Review of draft-yegin-dna-l2-hints-01.txt
> 
> OK, this is a good angle to discuss. I think another cause of 
> difference is about where we stand when we look at the L2->L3 
> info exchange. 
> 
> > So while an L2 Indication is what is a message containing 
> information 
> > authoritatively known by the originating Link-layer,  it is 
> called a 
> > Link-Layer Hint upon its transmission to the Network-Layer.
> 
> This at best can be "upon processing by DNA module". I think 
> downgrade from indication (or trigger) to hint would be 
> specific to the DNA. Some other consumer of such L2 
> information might not have the same usage of this 
> information. And we rather establish a common framework, 
> unless we want to see the same thing called differently among 
> different technologies (DNA, MOBOPTS, MIPSHOP, etc.).
> 
> I think it is perfectly fine to say "triggers/indications and 
> the associated information is used as a hint by DNA" in the 
> DNA solutions draft (and explain how exactly they are used).
> 
> Before the terminology gets all tangled, here is what I think:
> - it is OK to call it an "L2 indication".
> - I'd say an indication is generated as a result of a L2 
> event. It can carry an event name (link up, link down for 
> DNA) and optionally some auxiliary data (IP address, L2 
> access point address, etc.)
> 
> For now it suffices to define the events and auxiliary data 
> types that are relevant to DNA. But in the future, some other 
> WG can extend this....
> 
> Does this makes sense?
> 
> > This is because although it represents an event at the 
> Link-Layer, the 
> > entire message may be incorrect in detecting "link" change for the 
> > Network-Layer.
> 
> This is a view from within DNA.... 
> 
> Alper
> 
> > 
> > The link here being like in "IPv6 link-local" or "link" 
> from RFC2461.  
> > (I'd not normally qualify this, but we're talking terminology).
> > 
> > regarding L3 Hints...
> > 
> > I think that we're not chartered to change anything within or above
> L3,
> > except the detection but certainly we'd be interested in knowing if 
> > there's a port of the work we're not covering which we ought to.
> > 
> > If there's a task which we need to do, lets get it discussed (and 
> > possibly Bar BoFed, talked over with AD's etc, as is
> appropriate).
> > 
> > 
> > > I still owe response on goals and will do that as best I 
> can I have
> >  > been on vacation now and prior to that travel.  I am 
> working early
> in
> >  > the a.m. mostly during my vacation and will try to provide goals 
> > response soon.
> > 
> > Thanks Jim,
> > 
> > Your efforts are appreciated, but we also appreciate the value of a 
> > good holiday.
> > 
> > Please enjoy yourself.
> > 
> > Greg
> 
> 
>