CircleID

While the technical challenges are there, it's the politics that's the real challenge.

And not only the politics, but also the practicalities. Even if you could design a much better Internet from scratch, would it be worth anyone's while to actually deploy it? It usually takes a "killer app." or a factor-of-N improvement to overcome deployment inertia. Most protocol designs simply don't admit a factor-of-N improvement: that's usually reserved for hardware improvements. (The factor-of-N increase in address space between IPv4 and IPv6 fails to impress, because the net practical benefits are fairly small.)

I agree with Brett - look at the problems there are with getting people to move to IPv6 and that has very clear advantages. I think you can only really build on top of the Internet at this moment in time.

That said, the Net never fails to amaze me. All you need is some system to make it easy to shift and then an amazing leap in technology to encourage people to bother. However whenever anyone says that they *plan* to make a particular change, that's when I become sceptical. That suggests that people's ideas are correct from the start — and they never are. For example, Google initially planned to make *all* its money from selling search technology to businesses for internal use.

Kieren

I agree with Brett - look at the problems there are with getting people to move to IPv6 and that has very clear advantages.

Why should anyone move to IPV6? What's wrong with IPV4?