Thus spake "Gert Doering" <gert@space.net>
On Sat, Nov 10, 2007 at 08:12:27AM -0800, Christian Huitema wrote:Implementing "dual stack" is not much harder than implementing "IPv6 only", so this is pretty much a no-brainer.I want to challenge that. How are you going to implement "dual stack" if no more IPv4 addresses are available? Which is going to be a fairly typical scenario for new networks, some years into the future.
It's trivial to _implement_ dual-stack, but it's going to get difficult to _deploy_ dual-stack in a few years.
The response I'm expecting -- so I'll save folks the effort -- is that people will dual-stack with RFC1918 and NAT on the v4 side. My question is (still) why that is any better than v6-only hosts with NAT-PT. All modern apps are version-agnostic, so they have to deal with the ugliness of NAT with either solution, and having v6-only hosts cuts the work required by leaf host/network admins in half.
S Stephen Sprunk "God does not play dice." --Albert Einstein CCIE #3723 "God is an inveterate gambler, and He throws theK5SSS dice at every possible opportunity." --Stephen Hawking