Re: [DNA-BOF] Detecting Network Attachment BoF description (rev 2)

[Alper Yegin <alper@docomolabs-usa.com>]

>>>> How about we call this "Link-layer only" handover?
>>>> Is there a better term?
>>> 
>>> It is possible for multiple VLANs to be present on the same link, so
>>> merely determining that the host is on the same link isn't really enough
>>> to know that the address is still valid.  The issue is whether the host is
>>> still on the same subnet.
>> 
>> 
>> I thought about that case too, but I assumed change in VLAN means change in
>> link. This really depends on how we define "link", and how VLAN change is
>> implemented. If the host needs to re-do link attachment procedures in order
>> to change its VLAN, this might as well appear as a link change.
>> 
>> A similar case is when the access router has multiple physical links and
>> advertises a different prefix on each, although uses the same link-layer
>> address and hence the same link-local IPv6 address. I think we could call
>> your example "virtual link change".
> 
> I'm not sure about the term.
> There's definitely a link change going on here,
> even if some NIC manufacturers of devices with
> VLANs share MAC addresses.
> In this sense, ambiguity of v6 Link-local addresses
> is something we have to be aware of.
> 
> I think it is a 'definite' link change rather
> than 'virtual' :)

I mean Bernard's example is "virtual link change" whereas my example is
"physical link change". (BTW, I'm making up these terms, not sure if they
existed before). 

Alper