ietf
[Top] [All Lists]

RE: IPv6: Past mistakes repeated?

2000-04-24 16:20:04
Dick St.Peters writes:

| I should probably just go back to lurking

No, these are fundamentally hard problems, and nobody has real answers.
What is interesting is that people haven't asked real questions either,
and yet have decided that the correct approach is IPv6.

| but ... my take on every
| house being multihomed was to imagine full local meshing - each house
| peering with its neighbors redundantly.

Well, this presupposes some infrastructure built especially for
that purpose, or some surprising business-to-consumer cordiality.
I can't imagine for example the cable people offering free (if
local) connectivity to everyone for this purpose -- there is the
issue of owning&operating copper/fibre/brouter gear.  Likewise,
I can't imagine ANY operator who would allow unconditional access
to local infrastructure, including people who do not pay bills,
or who do destructive things.

| If, say, my power-line port
| was down, that information needn't be known by anything outside my own
| neighborhood.

Well, no - there is the issue of how traffic from elsewhere
in the world arrives to your neighbours' houses on your behalf.

That is, whatever arrangements were in place for the 
mapping 
your.place.ny.us -> 1.2.3.4/32 -> sprint -> netheaven -> DSL-line -> your router
would have to be adjusted to now prefer 
your.place.ny.us -> ... -> access -> your neighbour -> something -> your router

The "..." piece is what's interesting.

If the DNS were less "sticky" your.place.ny.us could map to a new address 
that already was going as far as your neighbour.

If the routing system were really robust to microscopic changes,
the DNS could stay the same and the difference would be that now
traffic would be attracted not to netheaven but to your neighbour.
This implies that the traffic is not attracted to netheaven and
possibly therefore also not sprint, but rather c&w and then "access",
which would now know that 1.2.3.4/32 goes to your neighbour rather
than to c&w.

| When the local power distribution center couldn't get a
| power-grid packet to me directly, they'd give it to my neighbor 

What if your neighbour is not being supplied by the power people,
because he doesn't like the price-performance and doesn't want
to pay the power people's bills?

What if your neighbour simply doesn't like you?

Why use your neighbour for backup when there is cable tv coax
running into your household also?

How can you deal with a failure at a power company transformer,
that cuts off both you AND your neighbour(s)?

| and
| let his smart house determine whether to send it to mine by wireless
| or cable or whatever else has come along.  The rest of the world could
| just engage in some kind of "get it closer" routing.

Right - the problem is that the physical geography and the network
topology can be disjoint.  Your neighbour may be physically next door,
and it may be that logically he is directly connected to you when
your copper through the wall is working.  But when that copper fails,
you have to go to netheaven, netheaven's next hop is sprint, sprint's
next hop is c&w, c&w's next hop is "access", and access finally delivers
over the cable plant to your neighbour.

I see no practical means of making sure that EVERY household
shares a single common logical supplier of traffic.

| Don't ask me about mobile users.  I'm going back to lurking ...

Don't give up - there is a learning opportunity for all of us.

        Sean.



<Prev in Thread] Current Thread [Next in Thread>