static vs dynamic addresses for apps [RE: v6 deployment in general [Re: tunnel broker deployment [RE: Tunneling scenarios and mechanisms evaluation]]]

[Pekka Savola <pekkas@netcore.fi>]

I'm responding to both Alain and Hesham in the same message, hoping to 
reduce overload on the list..

[Alain:]
> I'm not sure I want to follow you here. If an application needs to
> be written to handle interface renumbering to benefit from IPv6 end
> to end connectivity, this is putting the bar fairly high.

Apps certainly shouldn't need to handle renumbering on the fly .. or 
so I would hope.  But they should not store information like "this is 
my static prefix, the next time I start, I will use it" either -- in 
each case, they should determine it independently.  I don't think 
that sounds like too high a bar.

[Alain quoting me:]
> > Yep, I know this -- unfortunately.  But there is little that can be
> > done to fix ISPs which want to enforce the users not to do Foo
> > (whatever Foo may be).  We can hardly fix the misbehaving ISP problem
> > at the IETF.. as you're probably well aware :)
>
> In this scenario, we are already trying to fix the issue of ISP not
> cooperating to deploy IPv6... so I do not think it is unreasonable
> to seek to fix the address volatility problem at the same time if it
> is possible.

It is entirely different when the ISP does not *care* about IPv6, and
when the ISP is trying to *force* the user into a usage model (e.g.,
by renumbering the addresses used on the fly).  We're addressing the
first problem.  There isn't much we can do if the ISP e.g., filters
protocol-41 or the tunnel server UDP port at it's edge, for example.

Having to deal with the case where the ISP is harassing the user,
e.g., by the on-the-fly address changes is a complex problem.  If we
get a workaround for that for free -- that's fine.  If not, I don't
think we should try too hard.  In the end, the IETF is inevitably
going to lose to the ISPs who want to "encourage" their users to buy
their premium service..

On Tue, 16 Mar 2004, Soliman Hesham wrote:
>  > On Mar 15, 2004, at 1:11 PM, Pekka Savola wrote:
>  > > So, you want to build a simple application.  Don't we
>  > > all.. :)  But I
>  > > don't think this is something we can guarantee even with native
>  > > access.  
> 
> => Yes we can ! Why not? 

I think you're missing the context here.  Would you build an 
application which would only work if the user's IPv6 address is 
static, and not for the others?

At least when I have coded apps, I have to design them to work with 
everyone :).

>  > ISPs just aren't offering static v4 addresses 
>  > today 
> 
> => Yes they are ! You just pay more for them.

That's true of course .. but what do you think the user will choose,
if given a dynamic (or reasonably static) prefix for free, or having
to pay 5$/mo extra for a prefix which is guaranteed to be static?

I want the prefixes to be static as much as you do -- but I have
difficult time visualizing static prefixes to be so widely used that
applications, which could be run on all kinds of networks, could be
able to make an assumption that *all* the prefixes are static.  
Robust apps, especially if aimed at home users, just simply can't do
that AFAICS (I'd hope I was wrong..).

-- 
Pekka Savola                 "You each name yourselves king, yet the
Netcore Oy                    kingdom bleeds."
Systems. Networks. Security. -- George R.R. Martin: A Clash of Kings