New Year's Day has come and gone, but this month I'm writing to strongly suggest that you consider a somewhat belated resolution: make 2009 the year that you start researching, playing with, and just generally getting ready for IPv6. Wait, don’t run away. I know you've been hearing "IPv6 is coming" for nearly two decades (so have I), but I'm convinced that there's good reason to believe that you'll find it smart to know something about IPv6, and soon.
As you may know, IPv6 is an alternate protocol to IPv4, the protocol that's currently the foundation upon which the Internet is built. IPv4's great, but it only supports four billion addresses maximum, and as more and more of the world attains levels of wealth that let it afford Internet-connected computers, cell phones, and networks, the number of available IPv4 addresses dwindles daily. Thus, one day we'll run out, and by that day we'll need a replacement for IPv4 with a larger address space. IPv6 is intended to fill that bill by offering 18,446,744,073,709,551,616 (18 quintillion) addresses (and yes, I know that IPv6 is a 128-bit addressing scheme, but it essentially acts much like a 64-bit addressing scheme), as well as a bunch of other features that'll make an IPv6-based Internet more scalable.
Now, all of you who are reading this, raise your hands if you're currently using IPv6. OK, let me count—yup, just as I thought, almost nobody's using it. So why spend your time now implementing some networking stuff that will be of no value for ages and ages, and that probably will change drastically before you need it anyway? That was my thinking too, until about nine months ago.
It's Closer than You Think
Apparently, that far-off day of reckoning, the day that the various Internet authorities hand out the very last unassigned address, isn't all that far off, after all. Geoff Huston, an Internet watcher, runs a script every day that tracks how many IPv4 addresses have been assigned and, on the basis of past growth rates, predicts the day that the IPv4 addresses will run out. As I write this, his script (the results of which you can find at http://www.potaroo.net/tools/ipv4/index.html) predicts 18 March 2011 as the last day the Internet stops growing.
Now, let's suppose that Geoff's prediction is off by two years (although I should mention that other folks have created credible predictions placing that date at the end of 2010); in that case, we're still talking four years away, max. Okay, you might say, then I'll just wait until everybody else is doing it and I'll do it then, right? I'm going to suggest that might not be the best decision career-wise, and here's why. I first got my network on the Internet back in 1991, mostly out of curiosity—I was connected 24x7 with a 14.4 kilobit dial-up line that I never hung up (local phone service was a flat $15 a month) and my WAN/LAN router was an OS/2 box. Most large organizations weren't on the Internet until the web really took off around 1995. At the beginning of '95, nobody was on the Internet; by late that year, it was "get wired or get fired"—every organization wanted to get on the Internet, and they wanted to do it now. People who already had some Internet savvy prospered; others didn't.
Get Started Now
How's that apply to you, now? My guess is that IPv6 adoption will look a lot like IPv4 adoption: it'll be an all-of-a-sudden deal. One day, your boss will read something in an airline magazine about how everyone's going IPv6, he'll freak out and come ask you IT folks, "What are we doing on IPv6?" Of course, as your boss has never even uttered the phrase "IPv6" in the past, a response framed in a dumbfounded look might seem reasonable and fair, as in fact would an entire room of IT folks wearing dumbfounded looks seem reasonable and fair. But think how much better the meeting might work out if your response were something along the lines of, "well, boss, we've been looking into it and it seems that 85 percent of our software will probably run fine in an IPv6 environment, and it appears that our routers are all fully capable of IPv6—we've just got to configure them to use it. The biggest problem will be our ISP—they're a bit behind the times and just keep shrugging their shoulders when I ask them when we can get a block of IPv6 addresses, but I think we could use these other guys…." And, not to be, well, callous, but now suppose you're the only IT person in the room with these answers. Guess who won't have to worry about getting downsized during the next corporate restructure?
Seriously—just start doing a bit of reading on IPv6. If you’ve got Windows Vista and Server 2008, you've already got OSs that can let you build some test IPv6 infrastructure. A little future-proofing never hurts, you know?