Re: [v6ops] Operational Consensus on deployment

[James Woodyatt <[e-mail]>]

No, actually— it never needs to be feasible to turn off IPv4.  Ordinary
people, the users of Internet applications, don't really see any need to
augment IPv4 with IPv6.  And they never will until the utility of the IPv4
Internet begins to degrade noticeably in comparison to the IPv6 Internet.

If people are never exposed to the idea that "using $APPLICATION is full of
lose because IPv4, use IPv6 instead," then people will never have any
incentive to stop using IPv4 and start using IPv6.  So long as IPv4
continues to be 1) an essential service and 2) functionally equivalent to
IPv6 at the level of application visibility, there is no actual need to
solve the IPv6 access problem, because application users don't actually
need it.

If operators want ordinary people to migrate from IPv4 to IPv6, they might
try offering IPv6-only service at discount prices— which would position
IPv6 as the protocol for second-class citizens who can't pay premium prices
for the first-class Internet that rich celebrities enjoy between bites of
caviar.  Or they can offer dual-stack service and provision their networks
so that CGN44 boxes are the peak-load IPv4 bottlenecks, which would
position IPv4 as the protocol for second-class citizens who can't afford to
upgrade to the latest gear.  Which of those strategies do we think provides
more incentive to upgrade to IPv6?