> From: Keith Moore <moore(_at_)cs(_dot_)utk(_dot_)edu>
IPv6's claimed big advantage - a bigger address space - turns out not to be an
advantage at all - at least in any stage much short of completely deployment.
IPv6 deployment is going to have to be driven by IPv6's *other* features, and
when you take bigger addresses out of the cost/benefit ration, I'm even more
dubious that the features that are left (autoconfiguration, etc) outweigh all
the costs and risks of IPv6 conversion.
It seems that you can postulate whatever level of IPv6 deployment you like (a
long stretch in itself, but just for the sake of argument, let's make it) -
5%, 10%, whatever - and there's still no mechanism to drive further
deployment.
Here's why:
>> if you have a site which has more hosts than it can get external IPv4
>> addresses for, then as long as there are considerable numbers of IPv4
>> hosts a site needs to interoperate with, *deploying IPv6 internally to
>> the site does the site basically no good at all*.
> I do think that the main incentive to deploy v6 will come from the need
> to communicate with global addresses to points *outside* of folks'
> internal networks.
Huh?
If those outside sites are running IPv4, deployment of IPv6 does the people
who deployed it basically no good at all over IPv4 NAT - because the
fundamental problem (of not having enough external addresses) is the same,
regardless of whether the internal protocol is IPv4 or IPv6.
Thus, the problems caused by that limitation (many of which you so well
articulated in a previous message, such as the need to go through a
rendezvous to set up translation state) will also be the same, regardless of
whether the internal protocol is IPv4 or IPv6.
> deploying IPv6 internally .. will of course do some good if the site
> has applications on internal hosts that need to communicate with
> external hosts using global addresses. if you're .. point [is] that
> there's little purpose in having your own IPv6 island
Deployment won't do any good if the people it's trying to communicate with
externally are running IPv4 - and I *don't* include only the cases where
there's a local island of IPv6. Here's the logic:
As long as a substantial portion of the Internet is running IPv4 only, any
site is going to have to have some mechanism to communicate with the IPv4
portion of the Internet. It may not be quite as good as native IPv6-IPv6 -
*but it has to be OK, or the company is toast*. As an important corollary,
there may be an incremental improvement in functionality with a pure
IPv6<->IPv6 mode, but that increment is inevitably going to be minimal.
And my new point is that whether the site is running IPv4 or IPv6 internally,
it won't make any big difference to how well that IPv4 interoperation
mechanism operates, since the fundamental problem is the same.
So, if i) a company has an acceptable mechanism to interoperate, ii) they
won't see any big improvement from IPv6<->IPv6 operation, and iii) there's no
advantage in the IPv4 ineroperation mechanism to be gained by deployment of
IPv6 internally - then where's the incentive to deploy?
Noel