Joe St Sauver
JOE at OREGON.UOREGON.EDU
Sun Jun 15 18:08:11 UTC 2003
The real question when it comes to IPv6 is "What happens if I ignore it?"
At one level, the obvious answer is "Nothing, the planets all continue to
spin." It is clearly still a unicast IPv4 world out there. At another level:
-- There are numerous free tunnel brokers out there; random users can and
will get tunnelled v6 connectivity if native v6 connectivity isn't
available. Some *might* assert that transcontinental/intercontinental
tunnels aren't the best thing for the network (from a global perspective)
for a variety of reasons...
-- There are various 6to4 gateways out there, and they, too, are getting used.
If you enable IPv6 on XP, for example, by default 6to4 gets turned up.
Some might assert that 6to4 gateways provided by <fill in the blank> might
not be the best thing for the network (from a global perspective) for a
variety of reasons...
-- While you ignore v6, some of your competitors aren't. They're getting
address space, they're getting their router issues worked through, they're
getting their DNS servers squared away, they're working out local
addressing plans and figuring out application support strategies. They're
getting experience before they need it for production requirements, and
while it is still non-embarassing to be "just coming up to speed."
-- Or consider IPv6 in higher education as an indication that IPv6 may be
becoming ripe for easy deployment/may be beginning to become "real". At
least for most of the ten or so measurement peers shown at
you'll see v6 performance that's roughly comparable to v4 performance. Yes,
this is across Internet2/Abilene, but at least for IPv6, Abilene has been
peering with commercial v6 entities when everyone's in the right spot(s)
and it makes sense to do so.
I understand that in some cases hardware choices make v6 deployment hard (if
not impossible), and that more v6-incompatible-hardware is unquestionably
getting purchased out there as I write this. I understand that lots of
applications still aren't ready. And I understand that customers aren't
beating down your door saying, "Hey, when can I get IPv6?" (just as they
probably haven't beaten down your door asking about IP multicast in a
commercially meaningful way).
On the other hand, I do believe that IPv6 is one of those technologies that
will creep up on folks (if they aren't careful) simply because it does NOT
require "top down" deployment to get a toehold, and the fact that ISPs
AREN'T paying attention to it.
For example, even if many folks carefully track/shape/police IPv4 P2P
traffic, or block cusotmer servers, I'd be willing to bet that most folks
pay zero attention to their current v6 traffic (whether native, tunnelled,
And there *ARE* v6 P2P applications out there (the most notable of which is
probably Microsoft's 3 degrees), although others are rumored to be porting
their P2P applications to run dual stack. [There's a nice post about 3 degrees
entitled "Operational experience with 3 degrees" by Christian Huitema at
http://dict.regex.info/ipv6/v6ops/current/0183.html if folks are interested.]
And if you aren't paying attention to "Teredo" as a technology, you might want
to see http://www.microsoft.com/windowsxp/pro/techinfo/administration/
p2p/overview.asp -- quoting from that piece,
Teredo, also known as IPv4 network address translator (NAT)
traversal for IPv6, is an IPv6/IPv4 transition technology that
provides address assignment and host-to-host automatic
tunneling for unicast IPv6 connectivity when IPv6/IPv4 hosts
are located behind one or multiple IPv4 NATs. To traverse
IPv4 NATs, IPv6 packets are sent as IPv4-based User
Datagram Protocol (UDP) messages. [continues...]
Joe St Sauver (joe at oregon.uoregon.edu)
University of Oregon Computing Center
More information about the NANOG