[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 18:37:15 EDT 2021



Owen DeLong wrote:
>
>> On Sep 12, 2021, at 14:02 , Michel Py <michel at arneill-py.sacramento.ca.us> wrote:
>>
>>> Owen DeLong wrote :
>>> 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.
>> What you call short-term is 20 years and counting. I eliminated IPX in one year, and it was the #1 protocol at the time. For the last 20 years, I have heard that IPv6 would become the dominant protocol in 2 years. For the last TWENTY years.
> You’re mischaracterizing my statement a bit here.
>
> With regards to pain, anyone forced to maintain IPv4 on a network with any sort of ongoing growth is suffering an ever increasing amount of pain and costs.
>
> OTOH, once you have deployed IPv6 on a similar network, IPv6 is considerably less painful and life would be much better if you could turn off IPv4 at that point. IPv4 is the major source of ongoing pain.

And its the protocol you actually need to use.
>
>>> I use more. Are you going to claim that my choice not to NAT is somehow invalid?
>> In an IPv4 world, yes. It's irresponsible.
> That’s sort of like saying that people who bought single-family houses in urban areas should be forced into high density housing because in an urban area, it’s irresponsible to have a single-family house with a yard.
>
> I don’t buy the argument in either case.

I am with you on this. Yes. its theoretically "irresponsible" but that 
is no justification to be high handed.

>>> You’ve admitted that there are valid reasons for at least 6 addresses per household.
>> No. This setup is a hodge-podge that has no reason to be. In all of my branches all across the USA, I use only one IP and the eight or sixteen that come automatically are wasted. Intentionally.
> Even if you consider one address per site, the math still works out that IPv4 comes up short when you combine the number of households with the number of businesses and then add in infrastructure, servers, etc.

The fraction of households that actually care enough about guip + larger 
fraction of businesses/services might actually still work out mathwise.

Infrastructure does not count, that it consumes guip is a convenience in 
more cases than not.

Servers can be consolidated, again convenience.

>
>>> ISDN was quite widely deployed and, in fact, is still in widespread use, just not for data for the most part.
>> Oh ? where ? I did have a few PRIs at $job[-1], but there were local to my closet; the back end was SIP.
> There are a LOT of PBXs that interconnect with their ILEC or other Telco via ISDN PRI. VOIP has not completely supplanted it just yet.

I believe the original point was referencing the fact that ISDN was 
intended to replace POTS, nationally. Instead it found a solid niche 
role among providers and enterprises for a time, currently eroding. I do 
not think you prefer that narrative for IPv6.
>
>> For a variety of reasons, it was easier to communicate with legacy equipment with PRI trunks, but they came out of an Asterisk server; was a lot cheaper to buy a few Quad-PRI PCI express cards than upgrading a legacy system.
> In many cases, that’s still true, but instead of bothering with the Asterisk server, they are simply doing PRI with the Telco.
>
> Owen
>
Anyone doing that feel free to contact somebody willing to do that with 
you over sip, you will get redundancy and flexibility and its almost 
always a whole lot cheaper.

Joe



More information about the ARIN-PPML mailing list