Re: (ngtrans) including ipv4 address space into ipv6 address space

[[e-mail]]

On Wed, 14 Nov 2001 02:00:30 GMT, "D. J. Bernstein" said:

> Speaking as an IPv4 administrator: _I_ don't have a problem. I have all
> the addresses I need. I can reach the entire Internet, and the entire
> Internet can reach me. I'm not going to waste my time setting up IPv6;
> everything useful is reachable through IPv4.

That's your right, of course.

Although there *is* one small flaw in the reasoning:

You *cant* reach the entire internet.  What you *can* reach is "the part
of the internet that you find useful at the current time".  This is very
different than "the entire Internet".

Mention the three letters 'N', 'A', and 'T' to Keith Moore ;)

> Nobody is going to do that. Consequently everything useful will
> _continue_ to be reachable through IPv4. In short, IPv6 is a failure.

Nobody is going to do that *until there's a reason to bother*.

How quickly do you think IPv6 would deploy if Gartner released a
nice 8.5x11 color glossy report that said IPv6 was The Next Big
Thing for business applications, because it allowed enough address
space so the problems with getting 2 NAT'ed networks to talk to
each other can be resolved?

How quickly do you think IPv6 would deploy if msn.com or aol.com
provided <insert glitzy new product> as an IPv6 application?

How quickly do you think IPv6 would deploy if Quake4 supported it?

> Perhaps _you_ have a problem with this situation. Perhaps you want IPv6
> to succeed. Your only hope is to make IPv6 work WITHOUT ANY EFFORT FROM
> THE IPV4 SYSTEM ADMINISTRATORS. You have to slip IPv6 support into the
> operating systems and applications and routers so that IPv6 works
> _automatically_ over the existing IPv4 infrastructure. Then you have to
> wait for everyone to upgrade.

This is the same sort of wishful thinking that says "the only hope for
computer security is to make it work without any effort from the user".

Oddly enough, a lot of sites manage to get their users to do things
like install anti-virus software and personal firewall software, even
though it requires *some* effort.

Now that Microsoft has shipped IPv6 in Windows XP, most of the major
players now have IPv6 in their current product (only players I can
think of that aren't are Apple's OSX and SGI's Irix).  Now, we *do*
still have to wait for the upgrade crank to turn before "most" people
have been upgraded to IPv6 capable, so there's still a 3-5 year wait
before it becomes *really* common.

On the other hand, RFC882 came out in November 1983.  The HOSTS.TXT is
known to have been in use at least until August 1990, and may have been
updated throught 1992.

And even today, when you add a new system to the Internet, you often have
to tell it it's IP address, it's netmask, it's default router, and the
IP address of a nameserver.  Hardly "without any effort".

But somehow, the Internet and DNS have not been classified as failures
due to the incredible amount of work involved to get connected.  Users
grumble, but they do it - they've at least mostly gotten used to the
idea that they will occasionally have to put in some effort to keep their
machines functioning (be it upgrading, or patching, or fixing a corrupted
registry, or finding somebody willing to do it for them).


-- 
				Valdis Kletnieks
				Operating Systems Analyst
				Virginia Tech