Hello Erik,

Erik Nordmark wrote:

> > The same care is going to be required for routing headers as would be
> > for more general encapsulation in this regard.
> >
> > For routing headers, all that is required is to check that the home address
> > is the next intermediate routing point after the care-of address.  If these
> > addresses were inserted into a longer sequence of intermediate routing
> > points, the same check would be sufficient _for the purposes of Mobile IPv6_!
> > The other parts of the routing path in the routing header would have to be
> > checked according to the rules of whatever policy was used to build up the
> > other parts of the routing path.
>
> I missing something: My assumed use case is that folks
> want to use routing headers so that
> nodes can express a routing header with R1, R2, R3, Dest
> while limiting certain traffic to only express "MIPv6 routing headers"
> i.e. where there is a single hop on the final destination.
> In such a case which filter rules would apply on the various nodes.

My understanding is that the Mobile IPv6 use of routing headers
SHOULD NOT be managed in such a way as to preclude other
uses.  There was some discussion on the mailing list about this
a couple of months ago, and consequently some revisions were
made to draft ...-15 to enable more general use of the routing
header.  Before, there wasn't much discussion about it, but every
time that the topic came up I think that the sense was to avoid
usurping the use of the routing header.  Compatible co-existence
was more the intention.


> But those checks will not disable some other general facility like routing
> headers.

And, neither should Mobile IPv6, according to my understanding.

> Having a decapsulating node have a mechanism for various protocols
> that use tunneling to specify what is acceptable to decapsulate
> (so that MIPv6, configured IPv6-in-IPv6 tunnels, etc can all specify
> what is acceptable) would make a lot of sense.

Indeed, we could have a separate encapsulation for every application.
Or, as I hope, we can carefully specify the use of the existing tools in a
way to maintain their generality while still getting the particular function
we need.

> > In this way, no crippling of the utility of the routing header would result.
> > On the other hand, I hope that my point can be understood that all such
>
> I missing the point.

Is it clearer now?

Regards,
Charlie P.


--------------------------------------------------------------------
IETF IPng Working Group Mailing List
IPng Home Page:                      http://playground.sun.com/ipng
FTP archive:                      ftp://playground.sun.com/pub/ipng
Direct all administrative requests to [EMAIL PROTECTED]
--------------------------------------------------------------------

Reply via email to