[ppml] Incentive to legacy address holders

[John Paul Morrison]

Indeed, IPv4 is a huge block to IPv6.  (Aside: it took the Web for 
people to RUN to adopt IPv4. If we didn't have IPv4 and the Internet 
instead evolved on top of a competing protocol, we wouldn't be having 
any discussion about replacing the protocol; mostly because competing 
protocols had bigger addresses, and also possibly because the Internet 
would not have been as fertile ground for developing things like the 
Web, which made it take off).

I'd like to see IPv6 replace IPv4 and think it will eventually, but IPv4 
could linger well past its usable life (NAT will guarantee this). It's 
not like buying a newer faster car, more like getting everyone to switch 
to driving on the opposite sides of the road. Dual stack? Great, so why 
am I going to invest in training and man-power to upgrade everything 
when IPv4 is just good enough? IPv6 is here today in about every 
enterprise /service provider OS and network device, but it's not enough 
to get the ball rolling. It's getting into consumer devices but it's not 
enough.
Until you can talk to everybody with IPv6, you might as well talk to 
nobody with IPv6, because IPv4 already works.

What can policy do? It can nudge things along, remind us all that it 
would be a nicer if we all moved on to v6, and to help encourage things: 
we'll waive the fees for the early adopters, or better yet, simply keep 
your addresses and use IPv6-compatible addressing, since v6 is a proper 
superset of v4.  Policy can also drop a lot of baggage about how IPv4 is 
so flawed that we need a clean slate and we need more paper work and 
more BS to get going. We have a lot of running code but not a lot of 
consensus, but I think over time it will become apparent that we just 
need an "IPv6" enable button on our existing systems to accomplish a 
migration, and then people can just carry on, business as usual, just 
with a lot more addresses.


Ted Mittelstaedt wrote:
> IPv4 is ALREADY a block to widespread adoption of IPv6.  If we had no
> IPv4 we would all run to adopt IPv6.
>
> Right now I don't thnk the legacy IPv4 holders are the largest part of
> the foot-dragging-IPv4 holders out there.  So singling them out right now
> may not be merited.  However if they ever become the bulk of the
> foot-draggers
> then they are going to have to be singled out.
>
>