[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: Follow-up work on NAT-PT
> On 2007-11-11 05:12, Christian Huitema wrote:
> > I don't think that we need to optimize for IPv6 only hosts.
>
> I've been repeatedly told that we need to, with "real estate"
> usually being mentioned in the same breath. Otherwise, I'd
> agree with you completely - dual stack everything, with
> off the shelf IPv4 NAPT where you have to accommodate lack of
> IPv4 addresses. But the IAB explicitly asked for a solution where
> this is not the case.
We don’t have a real estate problem in PC-class machines, or even in PDA, smart phones and similar.
We do have a real estate problem in home routers, where we are witnessing a race to the bottom, but it is pretty hard to imagine home routers not supporting IPv4. So, if they support IPv6 at all, they will most certainly be dual stack.
I can imagine classes of devices that decide to go IPv6 only, but these will typically be single application devices, with no requirement to support IPv4.
In any case, real-estate arguments tend to be very speculative. The implementation size of any protocol can vary widely, based for example on trade-off between code size and performance.
What is the reasoning of the IAB? When they mention this requirement, do they have specific classes of devices in mind?
-- Christian Huitema