Re: [DNA] Plz clarify from draft-ietf-dna-link-information-00.txt

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

Hi Rajesh,

This seems like it's an interesting candidate to discuss with
respect to the multiple interfaces we've been looking at
in another thread.

Rajesh George wrote:
> Hi all
> 
> I am not clear about about sec 2.1 of the draft
> I have the following case in mind.
> 
> Case:-
> 	An MS already has an on going session with a WLAN and I suppose
> MS does not have GPRS coverage in that area. Now MS move out of WLAN
> coverage to GPRS area which does not have WLAN coverage. Now regarding
> the LL events
> 
> My confusion:- 
> 	 PDP context activation is triggered by the upper layer. As the
> WLAN session is going on upper layers are using L2 of WLAN as the
> underlying mechanism for transporting packets. Now how MS can initiate a
> PDP context activation?

In IPv6, the WLAN is likely to have IP layer connectivity of its own
without using GTP. So there will not need to be a PDP context on the
WLAN interface, unless GTP is being used as a global mobility signalling 
protocol (an alternative would be to use Mobile IPv6).

DNA is (typically :) run per interface, so movement out of WLAN coverage
may provide DNA with connectivity indications out of the link-layer
(Link-Down of the WLAN.

PDP context activation occurs to establish a link-layer connection
to the GGSN, which acts as a router on the GPRS link.   Upon
establishment of the context, an event notification (Link-Up) may
be generated (including IP addressing).

The establishment of PDP context occurs due to link-layer action,
though, not necessarily through network layer action.
When a host notices it has lost connectivity in WLAN it may bring up
its GPRS interface, though, which could initiate it (these decisions
seem to be out of scope for DNA though).

There's normally no need to link the WLAN to the GPRS unless the
address used for both is going to be the same (through some sort
of interworking).

Greg