[Leaplist] Good read. -- "Unfortunately, Someone Cannot Be Told What the IPv6 Is..."

[Bryan J. Smith]

Austin Denyer (Ozz) wrote:
> What would this mean for existing hardware, such as Cisco switches?
> For example, I have a Cisco 2924 and a 2820 - would they be trash with
> IPv6?

In the past, I've attempted to try to tell people not only how IPv6
works, but how it will work both in private networks as well between
AS on the Internet.  I don't just get a lot of dumb stares, I get
people who spread utter FUD because they don't know the first thing
about IPv6's design -- which was designs for coherence with IPv4.  And
they argue with me about things that have nothing to do with technical
reality, but their assumptions.

So, stealing a line from the Matrix, "Unfortunately, Someone Cannot Be
Told What the IPv6 Is ..."  I recommend you read up on IPv6, from
LINKLOCAL addresses to IPv6-to-IPv4 address translation, etc...  And
when you've think you've learned enough, read more -- especially real
implementation details.  That's the key.

[ In a nutshell, IPv4 networks can talk through IPv6 and vice-versa.
And LINKLOCAL in IPv6 is explicit, which is what many distros enable
out-of-the-box today.  I've long argued that everyone should be tested
on LINKLOCAL issues and debugging. ]

Long story short, over on the LPI lists, I was utterly disgusted.  I
like the organization and a few people involved, but when it came to
writing objectives, I got knocked on my ass by the "ignorant majority"
who have never worked on large scale enterprise networks outside of
web.

Now looking at in reverse, I've been in slammed with a lot of
consulting work, especially since I came here to New England (Long
Island Sound area).  So as of May, I really shrugged LPI off out of
sheer time after the April kick-off at LinuxWorld Boston.  I'm glad
they got the LPIC-3 exams done (thanx in no small part to Matt, he was
really the man "who got it done") as Japan was a major driver.

In fact, I'm kinda glad I "stayed out of it" because my objectives
would have thrown a major wrench into a very time-constrained
schedule.  There will always be time to "make it better" later on.

> I'd love to read that...

The concepts are extensive.  I've honestly had it with people looking
at layer-2 and layer-3 addressing and naming separate.  It's time to
solve the problem once-and-for-all with a solid framework designed
*ONLY* for private networks.  We can add features in from there (such
as the various keying/auth for DDNS, zone signing, etc...).