Re: [802.21] [DNA] Prefix information for link identification in DNA
Hi Mike,
On Mon, Oct 03, 2005 at 01:34:47PM +0000, Mike Moreton wrote:
> Yoshihiro Ohba,
>
> > Questions: What is benefit of having a generized definition of link?
>
> Well it means you don't have to remember all the specific definitions of "layer 2 link", "layer 3 link" etc. I'm lazy - that's why I like generalised, clean architectures...
>
Such a generalized term could be useful only when 802.21 really needs
to deal with all of those "links" at different layers.
> > In terms of handover decision making, should we define "link* up"
> > events for a "link" between a UE and an MIH Information Server that
> > may be tens of IP hops away from the UE?
>
> It depends what service you want to run over the link. When the link is capable of transferring the PDUs of the service running over it, then it's up. If you're saying there is no practical way for the UE to know whether such a link is up, then you've uncovered a problem with the protocol, not with the terminology.
>
> But surely it's the L2 and L3 links that you want to handover, not the link between a UE and a MIH Information Server? I don't really understand the 802.21 terms, so I may well be mistaken...
I agree that it's not the link between a UE and a MIH Information that
we want to handover. However, I think it's not L3 links either that
we choose based on the link events defined in 802.21 when we want to
handover (though we might need L3 or higher layer information that is
associated with each L2 link in order to choose an appropriate L2
link). Having said that, I believe 802.21 can define its architecture
with using the term "link" only for L2, and that seems cleaner to me.
(But probably this is somewhat a philosophical debate..)
Best regards,
Yoshihiro Ohba
>
> Mike
>