[6bone] RE: Content of 6bone digest, Vol 1 #446

[Douglas Wade Needham]

Quoting Jonathan Guthrie (jguthrie@brokersys.com):
> I have had other problems with SBC and I know that they really aren't
> able to handle anything but mass market "any color you want, as long
> as it's black" service.  You're probably doomed to disappointement if
> you're trying to convince SBC to do anything "leading edge".

And unfortunately, given how big SBC is and how they are used to
provide connectivity for folks like me, IPv6 is going to be tunneled
for folks like me for quite a while.

> > > before IPv4 goes away in the backbone.  Makes for an interesting
> > > conversation with my friends who are still working over at UUNET and
> > > AOL.
>  
> > I think the backbone is exactly where IPv4 CAN go away most easily.
> 
> I've got $100 US in my pocket that says that there will be at least one
> "backbone" provider with native IPv4 service 30 years from now.  (I
> would have said 100 years from now, but it's unlikely in the extreme
> that I'll live long enough to collect on that bet.)  Even if IPv6 is
> adopted vastly faster than I expect, IPv4 will likely not go away
> within the lifetimes of any of the participants here.  Perhaps never.

I would have to agree that IPv4 has quite a lifetime ahead of it.  And
while quite a bit of the equipment which was bought for Y2K has
depreciated, companies tend not to buy new equipment unless they
absolutely have to do so.  This is how Lucent has managed to layoff
the 70% or more of us who were working there Jan 1, 2000, including
myself.  They are still selling what amounts to legacy equipment with
just a few new features tacked on top of it.  Indeed, the 5ESS telco
switch which we designed probably 20 or more years ago had a project
to produce a replacement.  I know about half the folks who were
working on the 7E when it was canceled about 2 years ago.  A number of
those folks ended up working on my project here in Columbus, OH, and
in the past 2 years, over 70% of us either were retired early or laid
off. 

There are three certain ways to force a company to upgrade equipment
or software, at least in the telco sector.  They are:

    a) Lack of parts
    b) Lack of capacity
    c) Lack of a key feature (e.g. being able to handle Jan 1, 2000
       properly).

Anything short of that, and getting somebody to upgrade becomes nearly
impossible.  This is the main way they manage to give you dialtone
every time you pick up the phone.

I have been looking at this whole area for a number of years now.
Probably the best thing to happen to get IPv6 going in NA is to have a
killer app encourage the deployment of IPv6 so that each cell phone
gets an IP.  One other possibility would be fiber to the premises
(FTTP), so that we end up with more folks wanting to access things at
home while they are out and about.  However...

> > Me?  My pet desire is to see a tiny IPv6 stack that could easily
> > be used in embedded devices (< 1MB of RAM).
> 
> Don't want much, do you.  The IPv6 module I use on my computers is 280k
> long.  IPv4 takes maybe 1/10th that much.  In fact, that, along with a
> complete lack of multihoming solutions unless you're a TLA, is what
> prevents the widespread adoption of IPv6.
> 
> Here's a question:  What end-user routers for DSL, cablemodem, or ISDN
> access support IPv6?  If you want IPv6 to be adopted in North America,
> make that list longer.

This is probably one of the real reasons it will take at least a few
years.  I know quite a bit of the HW is capable of it, but the
companies writing the SW do not see a good ROI, so they are not doing
anything to get IPv6 support into their equipment.  You could say this
is a chicken and egg scenario.

- Doug

-- 
Douglas Wade Needham - KA8ZRT        UN*X Consultant & UW/BSD kernel programmer
Email:  cinnion @ ka8zrt . com       http://cinnion.ka8zrt.com
Disclaimer: My opinions are my own.  Since I don't want them, why
            should my employer, or anybody else for that matter!