On 9/15/21 4:26 PM, Owen DeLong wrote:
On Sep 15, 2021, at 16:20 , Michael Thomas <m...@mtcc.com
<mailto:m...@mtcc.com>> wrote:
On 9/14/21 12:44 AM, Eliot Lear wrote:
There were four proposals for the IPng:
* NIMROD, PIP, SIP, and TUBA
SIP was the one that was chosen, supported by endpoint manufacturers
such as Sun and SGI, and it was the MOST compatible. Operators and
router manufacturers at the time pushed TUBA, which was considerably
less compatible with the concepts used in v4 because of variable
length addressing. If we endpoints had some notion that v6 would
take as long as it has to diffuse, perhaps we all might have thought
differently. I don't know.
So I'm beginning to think that the reason ipv6 didn't take off is one
simple thing: time. All of the infighting took years and by then that
ship had long sailed. The basic mechanisms for v6 for hosts were not
complicated and all of the second system syndrome fluff could be
mostly be ignored or implemented when it actually made sense. If this
had been settled within a year instead of five, there may have been a
chance especially since specialized hardware was either nonexistent
or just coming on the scene. I mean, Kalpana was still pretty new
when a lot of this was being first discussed from what I can tell.
Maybe somebody else knows when hardware routing came on the scene but
there was still lots of software forwarding planes when I started at
Cisco in 1998 just as broadband was starting to flow.
Most of it was settled fairly quickly, actually. The bigger delays
were software vendors, network infrastructure product vendors (DSLAMs
and the like), etc. who even after it was well settled simply didn’t
feel a need to incorporate it into their products until about a year
after IANA runout.
I was aware of ipv6 -- or at least what would become ipv6 -- in either
1992 and absolutely no later than 1993. The internet was still tiny then
and I don't even think that CIDR was even a thing yet, and broadband was
still 5 years away. Had something simple like ipv6 headers, AAAA, and
the silliness of SLAC vs. DHCP wars had been resolved quickly there was
a very good chance that vendors would have adopted it if customers
wanted it or could be cajoled into wanting it. At that time IP was still
"the path to OSI" iirc so none of this was in cement from a vendor
standpoint (and sorry, there are lots more vendors than just router
vendors). But the entire thing dragged on and on and on and by then it
was far too late. Then the invention of NAT sealed that fate.
The IETF was a victim of its own dysfunction, film at 11 and now
we're having a 30 year reunion.
I’m not sure we can put all (or even most) of the blame on IETF
dysfunction here. Don’t get me wrong, IMHO there’s plenty of IETF
dysfunction and it is partially responsible. However, I suspect that
if IETF had rolled out the model of perfection and an ideal protocol 1
month after the IPNG working group started, we’d still be pretty much
where we are today because of the procrastination model of addressing
major transitions that is baked into human nature.
No I think the best was the enemy of the good. As it ever were. Nobody
seems to have appreciated what the real problem was -- time. Hindsight
is 20/20 but it wasn't hard to see that the internet was exploding and
wasn't in fact the "path to OSI" and by ignoring time there wasn't going
to be a path to anything.
Mike
