Re: [v6ops] Operational Consensus on deployment

[Ca By <[e-mail]>]

Allow me to posit that the locus of internet growth is cloud, IoT, and
mobile. Growth areas are important because they drive ip address demand.
Public IPv4 as a unique host ID is dead and will never be used natively in
these high growth edge networks.

Cloud is near 100% nat, the public address is abstracted from the host.
Look at AWS, they dont have a product that gives machines a public ip
address from AWS of any sort at any cost. It is all NAT.  My guess and
Azure and Google are similar, but i dont know.

IoT -- simply cant scale on ipv4.

Mobile -- always has been ipv4 nat in most places, and now approaching all
places.

Even in old world DSL broadband like AT&T DSL, layer 2.75 is private ipv4
that moves the 6RD end to end on top.

The bigger question is -- who thinks real  public dual-stack on every node
is viable for the timescale the ietf or edge network engineering operates
at?  Not T-Mobile US, not facebook, not at&t , not Amazon AWS, not
Terastream. All these networks run at least 1 stack virtually (or
relayed/proxied) and i doubt they see a path to real dual-stack.

The better ietf i-d i would like to read is a post-mortem on why dual-stack
failed in edge networks and what the ietf can learn from this failure.  It
likely has an economic and psychological theme.

Don't get me wrong, the utopian dream of everyone turning ipv6 on over some
period of time and ipv4 slipping off unnoticed into the sunset is great...
But clearly just a dream. It did not and will not happen.

The history of ipv6 transition will not involve a distinct period of time
where a public ipv4 and ipv6 dual-stack internet happened.

CB