The Internet is running out of addresses!
When devices are connected to our global network, each one is assigned a unique identifier. That 32-bit number is an IP (Internet protocol) address, and there are only about four billion possible numbers that can be stored in 32 bits.
Now, we find that there are only about six percent of those numbers left, according to a new piece in ReadWriteWeb. John Curran, an official from the non-profit that manages distributing IP addresses in North America, expects that those approximately 240 million addresses will be gone by the end of the year.
There is a solution: Internet service providers can switch to a revised Internet Protocol (IPv6) that uses a greater number of bits per address, thereby increasing the number of unique addresses by orders of magnitude. Experts, though, are sounding some alarm bells because that solution has to get deployed to work, and deployment of IPv6 has been slow.
What hasn't been slow is the number of devices being connected to the Internet, a fact which is blamed for exacerbating the address problem. "There's an explosion of data about to happen to the Web - thanks largely to sensor data, smart grids, RFID and other Internet of Things data," wrote ReadWriteWeb's Richard MacManus. "Other reasons include the increase in mobile devices connecting to the Internet and the annual growth in user-generated content on the Web."
Reading the Wikipedia entry on IPv4 address exhaustion, you'd probably believe that we're headed for some serious trouble as analysts try to predict whether enough people will switch over to the new IP language before addresses run out. But the recent behavior of companies is only a murky indicator of their future intentions. As ReadWriteWeb suggests, it's certainly possible that people will freak out when the moment of truth comes, and then quickly fix the problem.
We do happen to have a great historical analogue for the IP address issue. In the late 1980s and early 1990s, people worried that we were running out of phone numbers. The problem is essentially the same: you only have so many unique slots, and those slots eventually run out as phone numbers proliferate. So, in 1993, we get this UPI story: "U.S. Running Out of Phone Numbers."
"Americans have nearly used up the 1 billion phone numbers provided in a plan adopted in 1947," the piece reads. The plan to fix the problem was to release more area codes and start ten-digit dialing, which did indeed happen.
Again, it was the newfangled machines being connected to the network that were at fault. "Phone company officials blame the problem on increased demand for numbers by people adding extra lines to their homes and offices for such devices as computers and fax machines," the UPI wrote. "Cellular phones also have made a dent."
But a 2003 report [doc] from the company that runs our phone numbering system, the North American Numbering Plan Administration (yes, there is such an institution), finds that it wasn't just more devices, but a host of social, political, and technical factors. The report fingered bad planning by Bell, all kinds of technological things, government access, the break-up of AT&T, the competitive practices of the telecom companies, regulatory missteps, and "consumer's [sic] insatiable desire for increasing communication."
The "number exhaustion" problem, as NANPA calls it, never went away, but fixes kept being found because the system was too valuable to let fall apart.
Perhaps we can take two lessons from this analogue, despite differences in the technology and time. First, call us optimistic, but even if Internet infrastructure people have to scramble, it seems unlikely that the system as a whole will be allowed to fall into disarray. Second, in looking for reasons for the problem, gadgets tend to get blamed when they are actually just the most visible new thing. There are undoubtedly a swarm of issues leading to IPv6 underdeployed. But that's actually good news because it means there will be plenty of ways to fix the problem when everyone swings into action.