On 2009-09-23 21:05, Olivier MJ Crepin-Leblond wrote: ... > > Is a dual stack IPv4-IPv6 likely to be more unstable than pure IPv4 or > pure IPv6? Apart from the software engineering principle that more of almost anything is less reliable, there is a specific problem that if your computer believes it has IPv6 connectivity, but actually doesn't, *and* if your DNS service returns AAAA records, then you will see long delays in connecting to dual-stack servers while the application tries IPv6, times out, and finally tries IPv4. But of course some sort of flag waving exercise such as deprecating RFC 791 would have no effect on this in the real world. Running out of IPv4 addresses, and discovering that both nested NAT and port-based address sharing are horrible, will do it for us. Brian _______________________________________________ Ietf@xxxxxxxx https://www.ietf.org/mailman/listinfo/ietf