Re: [v6ops] Is there a problem? [was: Why enterprises aren't adopting IPv6]

[Owen DeLong <[e-mail]>]

> On Oct 1, 2021, at 13:32, Philip Homburg <[e-mail]> wrote:
> 
> 
>> 
>> Another way to describe it is what economists call "externalities", i.e.
>> costs that an enterprise creates for the community that it does not
>> pay for. The classical example is pollution - my fire pollutes your air,
>> and you pay the price while I use the heat.
>> 
>> The solution is "polluter pays", some way to force the cost of the
>> externality back onto the enterprise that creates it. How can we
>> force the costs back onto IPv4-only enterprises?
> 
> So it seems that you saying that running IPv6 has no net benefit in a 
> world dominated by IPv4. It seems that by now, eyeball providers have found
> some benefit in providing IPv6. 

Eyeball providers are out of IPv4 addresses and acquisition cost exceeds customer value. Eyeball providers need a much larger supply of IPv4 addresses to support IPv4-only content providers than the content providers do. 


Content providers are not (yet) facing shortage and their revenue per address is a much higher ratio than eyeball providers. 

This, like toxic polluters, lazy content providers are externalizing the cost of their laziness onto the eyeball providers. 

> However, we see that, for example, many content providers see no benefit in
> IPv6. Why is it that for them, a direct connection over IPv6 is not worth the
> effort compared to a connection through a CGNAT? 

Because they don’t pay the costs of that CGNAT. They create the cost, but the eyeball providers are forced to pay it. 

> Obviously, there is no way to force anybody to pay anything. We spend decades
> to make sure that internet traffic is cheap.

Major eyeball providers could, actually start demanding higher settlements for IPv4 peering than v6 and use that to start pushing some of the V4 costs back onto the content providers. 

> What can be done is to try find ways to make IPv4-as-a-service cheaper. 

That’s exactly the opposite of a good outcome, TBH. 

Owen

> 
> _______________________________________________
> v6ops mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/v6ops