A decade ago–earlier than that, in fact–there was much fretting that we would soon run out of IP (Internet Protocol) addresses, which are needed to connect computers and other devices to the Internet. It’s 2016 and we’re still using the same IP standard–IPv4–as we did back then. What gives?

The most obvious question is: why does this matter? Without IP addresses, nothing can connect to the Internet. No Internet, no email, no chat, no messaging, you get the idea. Addresses are essential. You can’t receive mail at home without an address, right? IP addresses are similar, but they are less permanent. The IP address your computer has right now may not be the address it has tomorrow, or next week, or next year.

Version 4 of the IP standard is what coincided with the Internet explosion of the 1990s, and this is the main reason why it remains in such wide use. It is deeply embedded in our infrastructure. Despite warnings about IP address exhaustion beginning almost as soon as the World Wide Web became a household phrase, and even though we’ve had IPv6 as a standard since 1998, we still rely heavily on IPv4.

A few things happened to reduce the urgency of a migration. Perhaps the biggest one was the transition to NAT, or network address translation. Back when running out of IPv4 addresses was a major concern, it was assumed every device connected to the Internet would need its own unique address. Since IPv4 provides only 3.7 billion public addresses (the rest are reserved for intranet and other special uses), there obviously aren’t enough for every person on the planet to have one. And with so many of us having multiple IP-enabled devices, like laptops, game consoles, tablets, and smartphones, it was only a matter of time before we’d run out. NAT held off this crisis by allowing users to hide a number of devices behind a single IP address. If you have a router at home, it is almost certainly using NAT to allow multiple devices onto the Internet. Your Internet provider sees only one IP address, even though you may have a dozen or more devices using the connection.

But even that could only hold off the inevitable. IPv4 was going to run out, sooner or later. And in 2011, that’s exactly what happened: the last few blocks of IPv4 addresses were allocated to regional Internet registrars. And yet, the Internet kept on hopping along. You haven’t suddenly found yourself unable to get online, right? It’s 5 years later and everything still works.

Part of the explanation is that IPv6 actually did get adopted, but in such a way that it’s invisible to most users. For instance, about 20% of the IP addresses in use in the US are IPv6, according to Akamai data. As long as most new address allocation is done on IPv6, v4 can stick around without too much trouble. Since companies and ISPs go out of business regularly, v4 blocks are freeing up on a regular basis, too, and get redistributed as needed. NAT is also still in wide use. There were also large amounts of addresses allocated to organizations which never used them, and over time a lot of that address space has been reclaimed and repurposed.

Interestingly, efforts to get all networks switched to IPv6 have been going on for over a decade, and yet adoption remains slow. The wide penetration of IPv4 as a universal addressing system has entrenched it, making a switch more difficult–people are unlikely to change something that already works, and IPv6 offers no tangible benefits to end users. IPv4 also has one convenience feature that IPv6 can’t match: it’s easy to memorize. A typical IPv4 addresses might be: 172.16.88.20. A typical IPv6 address, on the other hand, looks more like this: 2001:db8:85a3::8a2e:370:7334. Pretty obvious which one would be easier to remember or type in, right?

There has also been some resistance to IPv6 because of its relative permanence. With an address space so large we could never possibly exhaust it (it’s 2 to the 128th power, if you want to put that into a calculator), every device can have a single, fixed IPv6 address that it will never have to change or give up. On an IPv4 network, such the one provided by your Internet service provider, IPv4 addresses are cycled regularly. This makes it harder (but admittedly not much harder) for your activities online to be traced. A permanent IP address that follows you everywhere is quite different from one you’ll have for maybe a month before it’s given to someone else. That thin veneer of anonymity matters to some people, too.

Another reason IPv6 adoption has been slow is because it took a long time for software and hardware vendors to implement it well. It was not until Windows 7 that IPv6 support was a standard, built-in part of Windows, and that came out in 2009. Given that a lot of PC users are slow to upgrade–about 10% of desktops and laptops are still on Windows XP–it can take a long time for such a standard feature to reach a majority of users. Now that about 70% of users are on Windows 7 or higher, IPv6 adoption may pick up its pace.

Even then, I wouldn’t expect IPv4 to ever go away, at least as long as we’re using any kind of IP at all. Old hardware and software that don’t support it will stick around for quite a few years yet, and it’s unlikely to be removed from operating systems anytime soon–IPv4 support is built into IPv6. Over time, it will become a smaller and smaller share of the Internet, and someday, a future generation will puzzle about how we used to have such a tiny number of Internet addresses. And in that future, there will be a little network somewhere, still plugging along on IPv4 because somebody forgot to change it over.

Photo by chris__reed