Frank,
Wednesday, April 10, 2002, 12:51:08 PM, Kastenholz, Frank wrote:
> Alex
> I was just talking about the host mobility side of
> things. The requirements say that the routing system
> need not address host-mobility because mobile-ip seems
> to adequately cover it (if we're wrong, we'd like to
> know -- hence my note :-)
> As to network mobility (where, loosely, we can assume a
> 1:1 relationship between networks and IP prefixes), it
> seems to me that if the network moves, then its prefix
> "appears someplace else" in the topology. That seems to
> be "routing" to me.
So, here you seem to assume (correct me if this is not so)
that such a prefix is assigned to the network and gets
announced from different spots in the topology as the net
moves around. If so, this would be too much of an assumption
about a specific architecture, because as I said before,
one could come up with an architecture where all addresses
are topologically significant and so our mobile network
would get a new prefix each time it moves to a new distinct
piece of topology.
Regarding host mobility, I do not see a reason why a specific
existing technology needs to constrain the requirements.
I mean, if a new architecture allows a better approach than
we have now--let's consider it. See below as well.
> Hence the requirement that the
> architecture be able to cope with it. Now there might be
> a whole lot of ways to deal with it, maybe as a part of
> the routing protocols, maybe as a separate protocol, maybe as something I
> haven't thought of.
Having said the above, I do not think that an architecture
must be required to provide _built-in_ mechanisms for any
type of mobility (at least this is the way it sounds to me).
Mind you, there should definitely be a good story for mobility,
but there's a line between such a story (which may include
mechanisms above routing) and actual mechanisms in the routing
system.
Alex
This archive was generated by hypermail 2b29 : Mon Aug 04 2003 - 04:10:04 EDT