Re: Open question and Critical dependencies

[Julien Laganier <Julien.Laganier@Sun.COM>]

On Tuesday 29 March 2005 09:53, Dave Crocker wrote:
> Brian,
>
> > > >  Forgive me, but I do not see what it is about shim6 that
> > > >  cannot work equally
> > > >  well for IPv4.
> >
> >  IPv4 doesn't have a 64 bit routing prefix.
> >  IPv4 doesn't have an interface identifier field.
> >  IPv4 doesn't have a flow label field.
> >  IPv4 doesn't have extension headers.
>
> I wasn't asking for a detailing of differences between 4 and 6.
>
> I was/am asking what it was about the shim work that requires
> taking advantage of features specific to 6.

Hi Dave,

FWIW, here are my answers to some of your questions:

(...)

> Does/will shim6 use the interface identifier field?  If so, how and
> why?

The size of the IID field (64 bits) make it possible to use strong 
crypto (HBA or CGA) without requiring infrastructure (e.g. a PKI), by 
embedding in the IID a mapping information (i.e. mapping from an IPv6 
address to a set of prefixes, or to a public key). 

With IPv4 it is not possible to embed in the address some mapping 
information, because there isn't sufficient room.

> Does/will shim6 rely on the flow label field?  Why?

The flow label field is a good means to embed host pair context 
identifier for shim6 in IPv6 packets.

> Does/will shim6 use extension headers?  Why?

That might be a simple and clean way to negotiate, at the IPv6 layer, 
the activation of shim6.

>
(...)
>

Thanks,

--julien