[BBLISA] anybody doing IPv6 for real operations?/possible presentation topic

[Dean Anderson]

On Thu, 11 Mar 2010, Tom Limoncelli wrote:
> Other people (we'll call these people "the successful ones") go to
> their boss and say, "There's one specific thing I want to do with
> IPv6. Here's why it will help the company.  I promise not to touch
> anything else."

Except that there is no killer IPV6 app or service.  There is no one
thing that anyone "just has to do". After 15 years of pie in sky, IPV6
didn't and won't deliver any feature not in IPv4, except wider address
space. So, What do you get with more address space? Nothing. End users
can't get addresses.

Because the ISPs don't support IPV6 natively, the only way to connect to
anyone else is to tunnel to someone.  Well, tunneling is going to be
slower no matter what you do.  So now you have exactly the same app as
will run on IPv4, but its slower and can go less places. Nice marketing 
slogan. Reminds me of that turtle commercial.

I hate to rain on the parade, but that's basically why IPV6 is a
failure. Even if it didn't fail to interoperate, break stuff, and then
not work without IPv4, not break IPv4, etc problems, it would still be a
failure in adoption. An orphan, because there is no incremental
transition path for the global network, and the global network isn't
going to shutdown for an "upgrade day".

Pushing out a merely wider address space has to be done by ISPs;  only
they can do that and they can only do that using the feature set
available in ISO packet and routing protocols: IS-IS (multiprotocol
routing [ISO, IPv4, IPV6, Novell, etc) and CLNS (multiprotocol packet
delivery).  OSPF can only route IPv4. There is an OSPF6, but most
routers don't support that now, and an OSPF requirement is that //all//
routers must run the same version of OSPF4 and same version of OSPF6.  
You can run OSPF4 and OSPF6, but to route using OSPF6 all routers have
to know about OPSF6, even CPE.  Roughly same problem with BGP.

The IETF routing protocols all suffer from "not invented here, we don't
interoperate with foreigners, ambiguous specifications that must be
followed rigidly", while ISO design was "we route opaque protocol
packets securely with everyone without ambiguity".  So you can route
IPv6 on IS-IS on routers that don't know anything about IPV6.  You can't
do that with OSPF or OSPF6.

After a tough look around, the only protocol that can run //right now//,
and interoperate with IPv4 is CLNS.  Fortunately, NetBSD still has the
CLNS protocol stack, and the socket api works with CLNS addresses.  
Thanks to OSF, that funded this development back in BSD 4.4.

When we run out of IP addresses, we will be able to NAT between OSI and
IPV4 to keep things glued together, and transition services like web
servers to OSI native without breaking IPV4.  There will be no
uncertainty whether your OSI stack works with your router vendor.

		--Dean

-- 
Av8 Internet   Prepared to pay a premium for better service?
www.av8.net         faster, more reliable, better service
617 256 5494