Terry Slattery, who had been around the block a few times before I knew there was a block, makes several keen observations here about how we might see IPv6 deployed early on, what with all the ballyhoo about IPv4 finally receiving 2 in the hat. In summary:
- ISPs hand out IPv6 addresses instead of IPv4. Minus 1 for increasing ISP support costs.
- NAT continues its vaunted role as IPv4 life support. Minus 1 because it won’t last forever.
- ISPs start charging more $ to go beyond the few IPs required to uplink. Plus 1 for likelihood. [Ethan adds, minus 1 for style.]
- v6 to v4 gateways allow v6 clients to talk to v4 applications. Plus 1 for already happening today. [Ethan adds, minus infinity for adding infrastructure complexity while letting developers off the hook, at least for a while.]
There’s a bit more in the article, but I was left with one question. Are the network engineers going to have to “make us go“, AGAIN? Is it going to be up to the packet pushers to keep a bunch of legacy junk working because developers, sysadmins, and the rest just can’t seem to do what it takes to allow IPv6 to function as it was designed to? Okay, maybe not all the sysadmins. But you know what I’m saying.
IPv6 is a paradigm shift. I get that. But it’s dawning on me that the engineers aren’t the ones with the biggest problems. Paving the highways to support IPv6 will prove to be the easy part. Retooling the cars to run over the highway – that’s going to be the hard part. And because it’s hard, the next skills we network engineers will really have to develop aren’t IPv6 skills as such. Instead, the mad skillz will be in making all that old IPv4 junk work in an IPv6 world.
I have never felt so good about something so horrible as prolonging IPv4′s life. I expect to be employed forever.