I’ve done countless media interviews about IPv6, and among the first questions is invariably, “What is the ‘killer app’ for IPv6?” The answer is: “There isn’t one.”
The term “killer application” is overused to the point of meaninglessness. Technologies that are brought to market as “revolutionary” almost always are not. (Remember the Segway hype?) Real killer apps come upon us quietly; we wake up one morning and wonder how we ever got along without e-mail, the Web, mobile phones, search engines, or buttered toast. We begin using new technologies cautiously, maybe a little suspiciously, and the best of them work their way into the heart of our everyday lives. Any technology that has not yet proven its indispensability is a killer app only in the swarming minds of marketers.
There’s really nothing special about IPv6. Lots of lessons learned from IPv4 went into or are going into IPv6: A bigger address space of course, better address management, better mobility, maybe better multicast, hopefully better security practices, a so-far-elusive chance for better multihoming. But mostly IPv6 just allows the growth of IP networking to continue. Even if another IP application is never invented, we still need IPv6 just to sustain the growth of what we already have.
And that’s the business case. It’s not nearly as exciting as ubiquitous Internet, mobile ad hoc networks, or flying cars, but there it is.
The IPv4 allotment rates, as examined in the last two posts, show clearly that the IANA pool will run out sometime in 2010. In the last post I suggested that more indeterminate factors are likely to make the actual date that IANA allotments stop sometime in 2009. RIR pools will last about a year past when the IANA pool runs dry, and LIR pools will last less than a year when the RIR pools run dry.
The question is: When does all this actually begin effecting consumers?
First of all, it’s important to note that nothing is going to break when new IPv4 addresses become unavailable. Everything that’s running IPv4 at the time the tap is closed will continue running just fine. Most expansions of existing infrastructure behind existing NATs will work up to some limit of scale, although any new NATs will still need new public addresses.
The first significant effects will be felt by anyone wanting to build new infrastructure or offer new services requiring globally routable IP addresses, and who would normally apply directly to their RIR for address space. So developing nations and greenfield network operators feel the pinch first. But even large organizations with IPv4 space hoarded away will not hold out long, if they expect to significantly expand existing services.
The good news is that the first to feel the IPv4 squeeze are also the ones best positioned to take early advantage of IPv6: Developing nations and greenfield network operators are building new infrastructure anyway, and can incorporate IPv6 into their design, evaluation, purchasing, and implementation decisions without the transition considerations required from operators of existing networks.
If you are waiting for an IPv6 business case that involves some profitable new service or compelling new application, you will probably continue waiting right up to the time the plug is pulled on the IPv4 supply. The IPv6 business case is about capacity and survival. The cool new stuff comes after we have networks that can accommodate it.
