Re: [DNA] Comments on draft-shim-dna-proactive-00.txt

[Gregory Daley <Greg.Daley@eng.monash.edu.au>]

Hi Bernard, 

----- Original Message -----
From: Bernard Aboba <aboba@internaut.com>
Date: Tuesday, March 2, 2004 0:16 am
Subject: Re: [DNA] Comments on draft-shim-dna-proactive-00.txt

> > And please note that the CARD protocol can deliver information 
> required for
> > IP configuration such as the network prefix. So if the new AP 
> belongs to a
> > different subnet, the information previously delivered by the 
> CARD protocol
> > can help expediate the IP configuration after handoff. It can do 
> more than
> > just telling whether the new AP belongs to the same subnet as 
> the old AP.
> 
> As you noted above the SSID is not uniquely tied to the prefix.  For
> example, within a single SSID, associated Stations may have different
> prefixes.  So an advertisement of an SSID <-> prefix mapping is 
> only a
> hint.

While I guess this isn't what's intended by the draft,
I'd guess that there is enough potential uncertainty associated
with the AP information->Routing Information mapping to 
treat the configuration information as a hint, rather than a direct
configuration.

Certainly in some environments, this hint is likely to be quite
reliable, but there's still work to be done to specify what an
adjacent AP's actual configuration is.

The other individual submission to make extensive use of AP
information essentially provides a more constrained problem set.
The devices in the APID draft are able to guess that the access 
points provide the same configuration for this particular wireless
node, since they are already reachable from the local L2 broadcast
domain.

Incomplete information (association to an AP not in the list)
still is only a hint of change, as is association with a particular
AP (depending on where the APID config came from ... (hosts?)).

In both cases, the configuration has to be checked.  

That may mean a single pair of messages direct to that router, but
I'd guess that this is fairly fast if the configuration was correct
(a likely case).

Greg