Whenever anyone tells me that the sky is falling, I get suspicious. IPv6 is no exception – according to the forecasts over a decade ago, we were supposed to run out of IP addresses years ago. I even was quoted for a magazine article back in 2002 saying IPv6 wasn’t going to happen anytime soon.
This time around, I’m a believer – the IPv6 transition is happening.
There are three reasons this transition is real:
1. Everyone is getting onto the Internet. And I mean everyone. We’ve saturated the US and most of Europe for a while now, but China and India are bringing new users in droves. The way that these people are getting onto the Internet include both desktop computers as well as mobile devices which has actually led to *billions* of new addresses needing access. Today, NAT still factors in the solution to make IPv4 work, but with such staggering numbers, we’re starting to test the practical limits. This is making ISPs migrate to IPv6 and address solving the matter of how to make home Internet users migrate with them. Burton Group’s recent report on IPv6 (“Changeover to IPv6: The Deadline Approaches”) cited broadband heavyweight Comcast as already having made the shift.
2. There are places to go. Major web sites (many of whom use NetScaler) have been making the shift to support IPv6 for years. This solves the “who goes first?” problem and more conservative organizations can make the shift knowing that someone else has already debugged the networking hardware. (Not that my networking equipment has any bugs… None… Ever…)
3. The infrastructure needs to be ready. “But wait,” you should ask, “my vendor has been telling me that it has been ready for years.” Well, yes, most vendors have been IPv6 ready for years. Absolutely true. The trickier question is whether that readiness was complete, well tested, and fast enough to be used as the primary addressing mechanism. Having been on the forefront of IPv6 performance, we’ve seen and heard all kinds of horror stories where IPv6 was turned on and our networking peers would fall over on a fraction of the traffic or half the features wouldn’t work. With customers like Amazon, Google, Yahoo, and most significant of the IPv6 users, the US Government, we saw a whole manner of challenges.
This time, however, the vendors are sufficiently close to ready. The NetScaler for instance is able to show nearly identical performance in either IPv4 or IPv6. Combined with the years of experience in customer deployments, we’re more confident than ever than all the pieces that need to be ready truly are.
The shift to a pure IPv6 Internet won’t happen overnight, but it is closer than ever. Expect it to go from the inside out – meaning that the core providers will make the first move and use IPv4 to IPv6 gateways to manage the edge. As that edge moves further and further out, those companies that make the move will
get to remain close to the center whereas those that don’t make the move will be pushed out to the edge.
If you haven’t started looking into IPv6, start your homework. You don’t need to flip the switch tomorrow, but plan on it soon. For your web site, heckle your NetScaler sales guy (I get to heckle them all the time – it’s fun!) to make sure you’re sufficiently up to date to get the value of our latest work in IPv6 performance while you’re at it.
The sky isn’t falling, but I can see the sun starting to set on IPv4. And this time, it’s for real.