The secret Y2K disaster—but this one will really happen: The IPocalypse

Note: CS has a lot of acronyms; run your mouse over any of them to see what they stand for. Google them if you want to learn more—most of them aren’t really important to what I’m saying here.

We have roughly 100 days until the end of the internet as we know it.

Not exactly, but it’s technically true.

Image explaining the structure of an IP address

IPv4 is in 32-bit/4-byte format. IPv6 will be 128-bit/16-byte format.

Here’s a little bit of history: websites are assigned numeric addresses, called IP addresses. But you never see them is because they’re masked by named URLs, which you type in. The current format of IP address, called IPv4, has been in use since 1981. IPv4 addresses are in a 4-byte format (one byte is any 8-digit binary number, meaning it can store 256 different values).

Here’s a problem: do the math, and you’ll see that IPv4 can only store 2564 unique addresses. That’s 4,294,967,296 unique addresses. Guess how many are already taken.

Over 4,130,000,000. We have about 5% to go before we’re out of internets. And like everything else on the internet, new IPs are being taken at an exponentially faster rate than they were 5 years ago. According to this Twitter account, we’ve only got about 100 days until they’re all used up.

Well, the solution was invented 12 years ago. In 1998, IPv6 was made, which can store 25616, or roughly four billion IPs per person on Earth. But it’s hardly been implemented by our dear internets; every major OS, including those used by servers and big companies has IPv6 installed. But a 2008 study by Google reported that less than 1% of the internet is using IPv6. Very few have changed over, presumably because the end of the road hasn’t been close.

This is all conjecture, but with 100 days left and 99% of the internet having to switch to IPv6, one can expect there will be a lot of problems. Ever heard of Murphy’s Law? It’s a good rule of thumb to adhere to, especially in the computing world. And no matter how prepared everyone is, 99% of the internet is (or will be) almost 4,300,000,000 websites. There will be problems. A lot of major websites will be prepared (Facebook has already switched to IPv6, for example), but quite a few of those little websites you love just might go down in the switch to IPv6. And what then? It gets worse when you consider that, on top of these websites possibly going down, new ones cannot be created as there is no IP to assign them.


This entire thing could have been avoided if ZOI was followed. It’s the heuristic rule in CS that you shouldn’t limit how much data can be stored unless you have a very good reason. In other words, why limit it to 2564 IP addresses? You might as well prepare for the end-game and give it no theoretical limit.

A good practical example is on your cell phone: ever been annoyed that you can only store so many numbers per person? It’s pretty common for phones to limit you to having two ‘mobile’ numbers per contact, for example. But you only need one name for them, because they’ve always only got one name. (Okay, that’s not always true, but you get the point.)

But this technology was created in the 80s. Supposing we can forgive it. Even be proud of it for lasting this long!


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s