You may type www.biv.com into your browser’s address bar, but as far as the Internet’s concerned, there’s no such address. When you press “enter,” your computer contacts a domain name system database, which sends it to the address you really want: 209.205.95.212.
Each website or other Internet site has its own numerical address, a 32-bit number from a total of 4.3 billion unique addresses. When the Internet was being designed in the mid-1970s that seemed like a comfortably large number. Vint Cerf said that, back then, the Internet was thought of as an experiment that would be superseded when something more advanced was needed.
But that short-term experiment never ended. And now we’ve nearly run out of unclaimed addresses to assign to new users. On January 1, there were approximately 496 million unassigned addresses, down from 721 million a year previous.
The fine print: networked devices like the computers in your office or your smartphone only need an address that’s unique to their internal network. So your computer and mine might have the same address with no conflict as long as they’re on different networks. Despite schemes like that to share addresses, though, the pool of new, available addresses is running low.
There are predictions that the Asia-Pacific region will run out of free addresses this summer, with Europe running out by late 2011 and the North American pool drying up in 2012.
The end of Internet addresses as we know them has been predictable for some time. Since the mid-1990s, the people in charge of these things have been promoting a replacement for the current addressing scheme: Internet Protocol version 4 (IPv4) with IPv6. (No, I don’t know what happened to version 5.)
IPv6 addresses might look like FE80:CD00:0000:0CDE:1257:0000:211E:729C. The longer address means that there are vastly more possible addresses: more than three followed by 38 zeroes – a number that doesn’t yet have a name, but is in the ballpark of the number of molecules in the galaxy.
So we’ve had a solution to this issue for over 15 years. Nevertheless, we’re not ready for a smooth transition to IPv6. IPv4 and IPv6 devices do not easily communicate with one another, though dual-protocol devices are available.
Windows systems since XP service pack 1 and Mac OS X have supported both the old and new addressing protocols. But server and router support has lagged behind with many organizations understandably reluctant to replace IPv4-only hardware.
As a result, although major websites such as Google and Facebook are supporting IPv6, only a tiny fraction of Internet traffic (an estimated 1/20th of a per cent as of last October) was using IPv6. With such minimal use, it’s been hard to make a case for installing new IPv6-capable hardware. Canadian IPv6 deployment has been estimated at around 8%, significantly lower than the rate in Japan and several European countries.
What to do? As an end user, there’s no need to panic (though expect a new cable or DSL modem in the next while). If you’ve got a business or organizational network with only a few addresses accessed across the Internet, you’re similarly OK – no matter how many users there are on your internal network – but you should start making plans for a transition.
The big Internet service providers and telecom companies, though, are going to be facing a crunch soon – especially with the huge growth in demand for Internet-capable mobile devices. June 8 has been announced as World IPv6 Day, a day for test runs of new systems (www.isoc.org/wp/worldipv6day). You can test your IPv6 readiness today by browsing to www.test-ipv6.com.
The transition to IPv6 is coming. It might be messy. But it won’t be the end of the Internet.
Alan Zisman (www.zisman.ca) is a Vancouver educator and computer specialist. His column appears weekly.
