Re: [DNA-BOF] Attachment Detection charter text for discussion

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

Hi Spencer,

Spencer Dawkins wrote:
> Hi,
> 
> I agree with JinHyeock. "IP subnet change" is useful, but it's a
> subset of what's useful - transport protocols care about any
> path change (especially in wireless networks), and the competing
> traffic being encountered on the second AP is different from
> the competing traffic on the first AP, just to name one
> consideration (I'm sure there are others).

In some IP subsystems, change of link-layer
attachment point will be required too.
for example, IP multicast clients with
asymmetric link-layer visibility (snooping)
may need to undertake some IP related signalling.

Attachment to the same subnet after a period
greater than the  a DAD retransmission may
require re-DAD in some networks (as I think Alper
mentioned before).

While this work should be left to the individual
subsystems, there's no need to actually perform
the network attachment detection and identification
separately.

Greg.