Yet another instance where presuming the ip address assigned to a
device to be permanent causes a problem. I don't (personally)
think we will get out of this mess until we divide the concept of
an address as a topological locator, and an address as a host
identifier.
Russ
On Fri, 8 Mar 2002 john.loughney@nokia.com wrote:
> Hi all,
>
> > "Kastenholz, Frank" <FKastenholz@unispherenetworks.com> writes:
> > > If one proposes an architecture that does host-mobility, great. If
> > > one proposes an architecture that does not do host mobility, that's
> > > just as great. The reason is that host mobility seems to be a solved
> > > problem (Mobile IP). We chose to require that the architecture solve
> > > the unsolved problem (mobile networks).
> >
> > From where I look, I understand that Mobile IP does not involve at all
> > the routing system, it puts boxes at various places at the leaves.
> > But, IMHO, more experience is needed before considering host mobility
> > a "solved problem".
>
> I agree. One difficulty with Mobile IP is that it forces the device
> to send to/from a new IP address (care-of-address). Since most signaling
> protocols have been designed to be end-to-end with static IP addresses,
> Moblle IP introduces some disruptions where either the signaling
> connections need to be re-established, or the state needs to be
> transfered to the new point of attachment (something being done in SeaMoby).
>
> Anyhow, I do have an action point to organize folks into looking into
> mobililty (or micromobility) routing. There have been many proposals,
> most involving some sort of host-routing. Due to my day job, progress
> hasn't gone far on this subject. Hopefully, after this IETF, we can
> put some documents together & I can ping the list for review.
>
> thanks,
> John
>
__________________________________
riw@cisco.com CCIE <>< Grace Alone
This archive was generated by hypermail 2b29 : Mon Aug 04 2003 - 04:10:04 EDT