> If there is a means for all hosts to have addresses that are reachable from > all other hosts (barring that a security policy is in place), will companies > renumber their internal networks to coincide with this addressing scheme?
given that they could easily derive IPv6 addresses from the current IPv4 numbers in use, I don't see why not. even in IPv4, renumbering is only difficult because we haven't really tried to develop good tools for doing it. > If we (the Internet community) used private addresses and NAT for all hosts > that do not want/need/require access from the Internet, would the addressing > problem be as much of a problem as it appears to be? there's a dubious assumption that any significant group of hosts does not want/need/require access from the Internet - and especially that you can separate the hosts that do need such access from ones that don't. > If we are as generous with the IPv6 addresses, how soon before we have > the same address problem? so long that by the time it happened, we wouldn't be using any of the same platforms, applications, or protocols. to the extent that the danger of IPv6 address exhaustion exists, it isn't from using them like we do IPv4 addresses. it requires doing things like assigning a separate IPv6 address to every process created, or every smart pill that someone swallows (daily) to diagnose potential medical problems, or every network-capable toy that gets included as a prize in a cereal box, AND using lots of delegation layers in routing. Keith
