Re: how mobile do we want to be

[Thierry Ernst <ernst@sfc.wide.ad.jp>]

Sorry, this mail has nothing to do with Shim6, but let me clarify that
people in different sides of the world have different perceptions of
the needs of (future) applications.

I don't know what "bell-heads" means and I don't agree with this
perception of the application needs. The applications mentioned below
are applications we use today, what I'm talking about are the
applications we would like to have tomorrow, provided network mechanisms
have been designed in a way that will allow them (I would call this
"planning"). A bunch of them would need L3 mobility, and would be
assuming L4 stability. Session restablishment would be considered as a
bug, not a feature, by next generation users.

We may not live in the same side of the planet which may explain
differences in our respective point of view; both are worth to be
considered by enhanced protocols.


Thierry



 On Wed, 23 Mar 2005 11:21:19 +0100
Iljitsch van Beijnum <iljitsch@muada.com> wrote:

> On 23-mrt-05, at 3:07, Thierry Ernst wrote:
> 
> >> You have to drill a bit deeper as mobile phones aren't an
> >application > so it's impossible to determine whether they need layer
> >4 stability on >  top of layer 3 mobility.
> 
> > I'm afraid I don't understand the comment, but if it's what I think,
> > I would answer that to me it's pretty clear they need layer 4
> > stability.
> 
> No it's not. The only application I would really want this for is 
> remote login, but with the right tools reestablishing such a session
> is  trivial.
> 
> It would be nice to keep a/v streams running when switching 
> infrastructures, but congestion issues make this very problematic.
> 
> I think someone else mentioned one other application that could use 
> layer 4 stability on top of layer 3 changes, but I can't find this. 
> Stuff like web and file transfer don't need this capability, and in 
> mail only downloading large attachments would be helped by it, but 
> changing the pop/imap protocols can address this just as easily.
> 
> I suspect that the "need" for mobility is 99% driven by bell-heads who
>  still think in terms of services that must provide a seamless 
> experience to the user. Guess what: the "user" is a big company 
> building complex software, and if they can handle stuff like this 
> themselves (for instance, by restarting file transfers after an 
> interruption) they'll be happy to do so.
> 
> > I think that the discussion is about putting the right thing in the
> > charter, so I don't think it's coming at the wrong moment - rather
> > at the right time, before the WG is approved.
> 
> My answer remains the same: "not now". What could go into the charter 
> about mobility that would be helpful at this point in time? (And "we 
> need mobility, you guys figure it out!" doesn't qualify as "helpful".)
>  The wg can always recharter after finishing some specified amount of 
> work.