The problem with solving all problems with NAT (PAT) is that it enforces the client-server paradigm. The solutions that attempt to work around this restriction will inevitably cost considerably more (TCO) than just making a move towards something that is more scaleable (IPv6).

IPv6 is not a magic panacea, but it is a viable, evolutionary step in the right direction. I realize that when I say peer to peer, most will immediately think about file sharing, but it is more than that. It is a networking methodology. By continuing to constrict ourselves finding additional hacks to extend IPv4, we only serve to limit the possibilities of the next "killer app".