[arin-ppml] Change of Use and ARIN (was: Re: AFRINIC And The Stability Of The Internet Number Registry System)

Joe Maimon jmaimon at chl.com
Sun Sep 12 18:36:54 EDT 2021



Owen DeLong wrote:
> The refusal to deploy 240/4 are mostly on the basis that it would take just as much code effort to do that as it would to put v6 on a box, with the exception that most boxes already have a v6 stack, so actually more effort, yet yielding substantially less gain.
>
> So people who understand math looked at it and said “more pain, less gain, why would we do that?”
>
> I know you don’t like that answer because for some reason, you prefer the ongoing pain of IPv4 vs. the small short-term pain of deploying IPv6, but there it is.
>
>
With apologies to all, I will dig in a bit on this.

There is no fathomable way that code-wise it is the same effort. And had 
the naysayers stepped out of the way, it would have been done already.

Appeals to authority dont impress me, I have heard too many of these 
blithering blitherers to pay any heed to their supposed expertise and 
non existent impartiality, and that runs the gamut from the technical to 
current events (which probably explains our differences on those 
subjects as well). Sadly, there is no shortcut, you must engage your own 
brain.

There were many smart folk who detailed various ways 240/4 could be 
useful without global deployment the way that IPv6 could not and has not.

I have already deployed IPv6, back in 2008 on this very workstation. It 
has done nothing but (occasionally?often?) slow me down or cause some 
other mystifying connectivity issues.

I can count on one hand with leftover fingers the number of times 
non-geek customers, accounts, associates, projects, whatever had the 
slightest interest in spending real money on deploying IPv6. For 
established non provider networks there is no market force and demand 
for its primacy.  Its an afterthought at best, some sort of vague 
internet life insurance policy. Its the first thing that gets turned off 
whenever there are issues and the last thing to get fixed.

The only way IPv6 has been gaining is by sneaking in under the radar 
onto eyeballs. Not a whole lot of enterprise dough there. That does not 
make for an optimistic timeline, but its one of the brightest rays of 
hope IPv6 has currently got.

Enterprises of surprisingly decent size do just fine with business 
broadband /29s and /28's.

The reality is that the internet prefers, nay requires, IPv4.

Because IPv6 is extra pain for minimal gain until universal deployment. 
By design.

Going with the classic 240/4 objection, it should never have been released.

But it avoids CGNAT! The users who care about that care more about a 
public IPv4 ip address, which by and large they can still get. Just like 
a lotto ticket, the first one is the real value, additional merely 
incremental.

The networks that care about that are the ones sneaking IPv6 into 
eyeball users, and the ones who have not done so have run the numbers 
and decided they dont care and you cant make them. Which we hear about 
quite often here and elsewhere.

 From the enterprise and user point of view, there was no first mover 
advantage to IPv6 deployment. There is very little present mover 
advantage. Instead, there is strong incentive to be among the last to 
spend any effort on IPv6 migration. And most folks have figured out that 
could be a ways off or possibly never.

Which circles back to the point that the only apparent real world 
advantage to network and service providers to their deployed IPv6 is as 
an offload optimization to nat/proxy layers.

Turns out that IPv6 was just a premature network optimization.

Joe


More information about the ARIN-PPML mailing list