Re: IPv6-only devices?

[Keith Moore <moore@cs.utk.edu>]

> It makes sense to deploy IPv6-only nodes and services in two
> situations:
> 
>          (1) Closed or limited network situations where you
>                  know that all nodes that need to reach the
>                  IPv6-only node/service have IPv6 available.
>
>          (2) New services/applications that require a
>                  software upgrade, anyway.  While the
>                  client is being upgraded to include the new
>                  service, IPv6 can be installed or enabled.

not that these are entirely disjoint, but there are other categories
worth mentioning.  e.g.:

(3) services/applications that will make heavy use of portions of the
network that are IPv6-only, or network nodes that are only
capable of IPv6

in which case it might make more sense to impose a burden on the
(presumably far fewer for such applications) nodes that are currently
limited to IPv4, to upgrade to v6 stacks and use 6to4 or tunneling or
some such to communicate with the other nodes.

(4) new services that involve adding large numbers of new
individually addressible hosts, for which there are not enough IPv4
addresses available

for instance, various kinds of devices that would benefit from remote
monitoring or sensing (power meters, traffic sensors, security
cameras/sensors, fire/smoke alarms, etc.).

> It may be a long time, though, before it will make sense to
> deploy IPv6-only servers on the public Internet to provide
> existing services.  For example, how long do you think it will
> be before it makes sense to deploy an IPv6-only e-commerce
> site?

Right, for services where there is a wide expectation that v4 access is
sufficient (email and the web being really obvious examples) you're
going to continue need to provide access using IPv4 for many more years.

Keith