RE: The state of IPv6 multihoming development

["Tony Li" <Tony.Li@procket.com>]

|   I wouldn't expect domestic users to be MH, so I don't expect to see
|   a million MH sites in those cities.


Brian,

I would ask that you sit and rethink this for a moment.  While being
MH right now is beyond the needs and costs of most people, I can see
a future in which this is not true.  With many folks attempting
to solve the last mile problem, I can see high speed bandwidth
available via fiber, cable, wireless and DSL in the next few years.

The competition will drive down cost.  And those folks who find
net access to be a requirement for their daily lives, such as
telecommuters, may well decide to MH.  We see this already with
certain netgeeks today.

Also, small businesses are also candidates for MH.  We already
see gas stations with satellite dishes.  Other folks are already
putting IP addresses on their cars.

Given all of this and the fact that we want an architecture that
lasts, we should have the flexibility to accomodate this future.
This level of MH has an interesting impact.  We need to be able
to support millions of MH sites in a "metro" area.  The operational
reality here is that flat routing at this level is probably not
a good idea.  This implies that if we were going down the
geo path, we would really want to do this at finer granularity.
I'd suggest that we need to be at something like CO granularity.
Of course, multiple levels of hierarchy would be beneficial, so
this is in addition to metro aggregation.

Make sense?

Tony