Shim6, was: Re: filtering /48 is going to be necessary

Mon Mar 12 19:07:12 UTC 2012

On Mon, Mar 12, 2012 at 11:31 AM, Leo Bicknell <bicknell at ufp.org> wrote:
> In a message written on Mon, Mar 12, 2012 at 11:07:54AM -0400, Robert E. Seastrom wrote:
>> Grass-roots, bottom-up policy process
>> +
>> Need for multihoming
>> +
>> Got tired of waiting
>> =
>> IPv6 PI
>
> It was never clear to me that even if it worked 100% as advertised that
> it would be cheaper / better in the global sense.

Hi Leo,

When I ran the numbers a few years ago, a route had a global cost
impact in the neighborhood of $8000/year. It's tough to make a case
that folks who need multihoming's reliability can't afford to put that
much into the system. As long as the system is largely restricted to
folks who do put that much in, there's really no "problem" with the
current flood-all-routers multihoming strategy: at $8k/year the demand
will never again exceed the supply.

A *working* multi-addressed end user system (like shim6 attempted)
could solve cheap multihoming. That could have a billion dollar a year
impact as folks at the leaf nodes decide they don't need the more
costly BGP multihoming. But that's not where the real money is.

Often overlooked is that multihoming through multi-addressing could
solve IP mobility too. Provider-agnostic and media-agnostic mobility
without levering off a "home" router. That's where the money is. Carry
your voip call uninterrupted from your home wifi on the cable modem to
your cell provider in the car to your employer's wired ethernet and
back. Keep your SSH sessions alive on the notebook as you travel from
home, to the airport, to London and to the hotel. Let folks access the
web server on your notebook as it travels from home, to the airport,
to Tokyo and back.

The capability doesn't exist today. The potential economic impact of
such a capability's creation is unbounded.

Unfortunately, shim6 didn't work in some of the boundary cases. Since
single-homing works pretty well in the ordinary case, there's not much
point to a multihoming protocol that fails to deliver all the boundary
cases. IIRC, the main problem was that they tried to bootstrap the
layer 3 to layer 2 mapping function instead of externally requesting
it. That's like trying to build ARP by making a unicast request to a
local router instead of a broadcast/multicast request on the LAN. What
happens when the local routers no longer have MAC addresses that you
know about? Fail.

Also, in complete fairness, shim6 suffered for the general lack of
consumer interest in IPv6 that persists even today. It's proponents
bought in to the hype that new work should focus on IPv6, and they
paid for it.

Regards,
Bill Herrin

-- 
William D. Herrin ................ herrin at dirtside.com  bill at herrin.us
3005 Crane Dr. ...................... Web: <http://bill.herrin.us/>
Falls Church, VA 22042-3004