Re: [Fwd: [Mobopts] New ID: Multicast Mobility in MIPv6: Problem Statement]

Hi Kevin,

thanks a lot for the comments:

Kevin C Almeroth wrote:
> Okay, I took a brief look...  so other than some work on the
> grammar (which I can help with if you need it), 

mhmm, we should be able to fix this ... but are very grateful for any 
pointer or hint.

> I think the
> major improvement would be to strengthen what you/we/mobopts
> thinks are the major things that need to be considered in
> designing a mobile multicast solution.
> 

Yes, you're right: there is no roadmap or working perspective, yet.
This is what we should discuss in the meeting next week: up until now we 
tried to identify/describe the problem and the solution space as clearly 
as possible, no conclusions from that.

> Some additional thoughts:
> 
> 1.  Multicast, as it turns out, has been tremendously successful.
> Not as a ubiquitous, public Internet IP service, but as a 
> private, enterprise-based service.   What this says is that
> individual wireless/mobile networks will need to provide wireless,
> but the need to connect to a global infrastructure may be somewhat
> limited.
> 
Yes and perhaps, I'd like to state ;) Interdomain ASM suffers from 
technical complexity and there is reasonable anxiety to promote DDOS by 
offering ASM beyond site limits. This may change with SSM and there are 
many attempts to establish group distribution services on the 
application layer ... so there is a need.

> 2.  One of the problems why mobile multicast hasn't taken off is
> that there isn't a need...  yet.  Furthermore, there really isn't
> really mobile networking taking place, at least not with seamless
> handovers.  Do you have any experience with anyone running a
> real MIPv6 deployment?  If such a deployment doesn't exist and
> seamless handovers don't exist, what is the need for seamless
> multicast handover?  

Yes, this is a hen/egg problem. But we are all optimistic to see 
seamless (F/M-)MIPv6 services spread ...
... but you're right: just returned from Cebit I saw tunnel-based 
mobility proposals/prototypes aiming just to avoid mobile IP.

> Maybe we should start with something
> simple, which is just a multicast solution that works in 
> wireless networks.  Sure, existing protocols work, but are
> there things that should be changed to better support multicast?
>

D'accord - this most likely is the attempt of a deployment-centric step 
following the problem statement.

> I'm sorry I won't be in Prague to see the presentation...  I
> think there is a real opportunity to do something valuable here
> (as opposed to a solution that looks good on paper but no one
> cares about or deploys).
> 

Thanks - we'll try our best ...

See you in Copenhagen then

thomas

_______________________________________________
Mobopts mailing list
Mobopts@irtf.org
https://www1.ietf.org/mailman/listinfo/mobopts