Re: Minutes / Notes

[Masataka Ohta <mohta@necom830.hpcl.titech.ac.jp>]

Kurt;

> >>>> b) Introducing loc / id separation will require mapping, one way or
> >>>> the
> >>>> other.
> >>>
> >>> Wrong. The separation requires that a host know id and locators of
> >>> its peer with reasonable security.
> >>
> >> That is a mapping state in it self.
> >
> > It is a state. You can call it a mapped state. But, there is no
> > mapping service required.
> 
> This depends on what you mean with mapping service.

OK, we have no disagreement here, save minor terminology issues.

> > That no service required means no additional security required.
> 
> You suggested that the tokens be transfered OOB,

No, not in general.

However, it is requied for mobility though.

It should be noted that, when IPv4 mobility was developed, security
experts reviewed the protocol and modified it to reuire the security
with the shared secret transfered OOB.

> that is a security and 
> a state as well. Most likely a very costly one.

So? Mobility costs, of course.

> >>>> This introduces new bindings that needs to be secured.
> >
> > is wrong and the existing bindings has certain security which
> > is just enough for weak security.
> 
> The bindings will need some form of security. How you do this is part 
> of the discussion.

I have been saying that that they are contained in a single packet
means binding with just enough security.

						Masataka Ohta