On Nov 22, 2012, at 10:37 PM, George Michaelson <g...@algebras.org> wrote: > On 23/11/2012, at 1:18 PM, Ted Lemon <ted.le...@nominum.com> wrote: > >> On Nov 22, 2012, at 8:46 PM, John Levine <jo...@taugh.com> wrote: >>> PS: If you were planning to say that with the magic of IPv6, everyone will >>> be able to run servers on their home cable connection, don't bother. >> >> Why not? >> > > Because the lack of public IPv4 was only an *excuse* for blocking this > functionality: ISPs in the main don't want you to equalise traffic coming > from their CAN network out to the world, they didn't build with that model in > mind.
I hear you, but this isn't really a meaningful rejoinder, because this equalization you are talking about isn't a prerequisite for selling devices that will want to publish services from the home. You can't have a network that only lets traffic go in one direction. IPv4+NAT makes it hard for traffic to go in both directions, and yet still we have bittorrent, Skype, limited VoIP, and lots of back-to-my-pc solutions. We have dyndns.org, which wouldn't have needed to exist if people weren't setting up home servers. So really, while it may be true that ISPs will resist certain kinds of home servers being set up, and I don't people to be running web servers targeted to the general public on home network connections, it's entirely reasonable to think that low-use home servers will be a reality in the near future. We know that ISPs are talking about serious, high-reliability M2M applications. So external reachability is going to be a reality. It's easy to dismiss this (which I don't think you were doing—I'm really referring back to what John Levine said) by putting down home servers as a "hobbyist" sort of thing, which ISPs won't feel market pressure to support, but I think that this attitude is actually what's unrealistic. There is a *huge* untapped market in peer-to-peer applications. We can think now about how to make sure that this market has the tools it needs to grow, or we can ignore it until it becomes big, at which point it will be too late. _______________________________________________ DNSOP mailing list DNSOP@ietf.org https://www.ietf.org/mailman/listinfo/dnsop
