This was published 13 years ago
Crunch time for upgrade of internet addresses that are running out
By Samantha Amjadali
With cyberspace almost full, Samantha Amjadali finds out it's going to take half a trillion dollars to avoid a global squeeze.
IPV6. It is the ultimate case of procrastination; a problem so big, so complex and so expensive, the world has ignored it for two decades.
IPv6 Forum Australia president Michael Biber (left) says businesses have been reticent to switch to Internet Protocol version 6, a situation that APNIC chief scientist Geoff Huston (right) calls a "textbook market failure".
The problem: the internet is full. Well, almost.
It is brimming with more than 4.29 billion websites and addresses for networks, mobile phones, refrigerators and, of course, personal computers at home and at work.
As with anything that exists on the internet, anything connected to the internet and anything transported across the internet, an address is required — called an IP (internet protocol) address — that acts like a phone number, allowing every individual device, piece of data or location to be found in the digital labyrinth.
But the existing 32-bit internet routing system, Internet Protocol version 4 (IPv4), is only capable of hosting 4,294,967,296 addresses. So, when the remaining global pool of some 204 million addresses runs out, as is expected before June, no new devices or websites will be able to connect — unless they've transitioned to IPv6, the latest 128-bit incarnation of Internet Protocol, which can store up to a whopping 340,282,366,920,938,463,463,374,607,431,768,211,456 (or 2^128) addresses.
"IPv4 is about 30 years old and when they put it together, they anticipated 4.3 billion addresses would last a long time," says Michael Biber, the president of IPv6 Forum Australia, the local chapter of the global IPv6 Forum charged with spearheading the transition.
"Of course, they couldn't have predicted what would happen with mobile phones and VoIP and social networking, or that we'd use up addresses as fast as we have."
Mr Biber says that though the IPv6 Forum has been working on the transition for more than a decade, businesses around the globe have been reticent — until now — to exploit their efforts.
The biggest issue is that IPv4 and IPv6 don't talk to each other. Though IPv6 devices communicate happily with their less-evolved IPv4 cousins, it doesn't work the other way.
A breakdown is inevitable. It won't be immediate and it won't be dramatic but it will happen, slowly and subtly, as little bits, mostly the newer IPv6 bits of the internet, fail to appear to anyone still using IPv4 architecture.
The IPv4 internet and all the devices and sites we're accessing now will still work exactly as they do now but when IPv6 devices and sites start appearing online, the existing IPv4 infrastructure simply won't be capable of recognising them and small, random pockets of the new internet will be inaccessible by older hand-helds, routers, modems, servers and operating systems.
Upgraded websites won't appear on machines running old software and hardware and even your shiny new IPv6-ready handset will not be able to display your weather update because your telco may not have upgraded its own IPv4 back end.
The solution
So, who's in charge of fixing this looming, global problem? Well, you are.
Nobody owns the internet and, technically, nobody runs it. It is a hotch-potch collection of billions of networks and connections and machines all based loosely on protocols defined by ad hoc bodies with no official power to do anything, really.
Most governments have delivered mandatory IPv6 upgrade dates for their public services and networks to be made IPv6-compatible but don't have the authority to force private or corporate citizens to upgrade.
That means all internet users have to take matters into their own hands and upgrade their own points of connection — from the ISPs who need to upgrade their servers to host IPv6-compatible websites and the vendors who need to upgrade their mobile handsets, to mums and dads who may require new modems, wireless routers and IPv6-compatible mobile handsets.
It's like the biggest team effort in the world.
Consumers will likely get away with the smallest dent to their hip pockets.
Thanks to the relatively short product cycle of today's tech gear, when your current hardware becomes too old or too slow, it's simply a matter of upgrading to an IPv6-capable device, which vendors such as NetComm, Billion and Cisco are already selling.
Of course, some hardware and software will be able to be upgraded via a software patch and many items such as Macs, Windows Vista and Windows 7, as well as many corporate-edge devices and any of Apple's iOS 4 devices, are already IPv6-capable. But, says Geoff Huston, the chief scientist for APNIC (Asia Pacific Network Information Centre), the biggest regional allocator of IP addresses globally, the main problem won't be at the consumer end.
"The problem will be the filters, the ISPs and all the bits in the middle that people don't see that aren't capable of running IPv6," he says. "It's the underlying infrastructure that will actually stop the larger numbers of people."
The only parties who stand to make a clear profit from the transition are the hardware vendors selling new IPv6-compatible gear. For everyone else, the upgrade will cost something but there won't be much to show for the investment.
It's everybody's problem but nobody is in charge and because there is no financial advantage to be gained from upgrading to IPv6, nobody really cares. "They're [telcos, ISPs and so on] all thinking that it will cost to upgrade to IPv6," Mr Huston says, "which it will, but I don't get any more money in return, I don't get any more customers and as long as my competitors don't upgrade, I won't either. That's basically what's been happening for the past decade. It's a textbook market failure."
Even with a conservative estimate, Mr Huston says the global costs for the IPv6 upgrade will exceed half a trillion dollars but acknowledges it's a nebulous figure, though he suspects even this staggering estimate may be on the low side.
"There are a number of ways to calculate the transitional costs for IPv6," Mr Huston says. "One way is to take the internet's 1.7 billion users globally and guesstimate about $100 per user for new hardware, which adds to $170 billion.
"I suspect that this is low but it's a tough one to attempt to quantify and I suspect that $200 per user may be closer when you factor in support and new hardware. But it will cost more for some and less for others.
"So, rounding up just a little, maybe half a trillion dollars for the transition globally."
IPv6 Down Under
In Australia, the internet service provider Internode has led the IPv6 transition, trialling IPv6 with customers for more than a year after deploying it internally in 2007.
Last month, it also launched the country's first sub-$100 IPv6 router for consumers and its IPv6 trial, which involved many of Internode's tech-savvy customers, has been where many vendors have been testing their hardware.
Competing ISPs have been watching, transfixed.
Globally, the smaller countries and even smaller ISPs are grabbing the IPv6 bull by the horns. The transition's leaders can be found in places such as Slovakia, where early government intervention has meant that all its ISPs are now IPv6-capable, as well as several African nations such as Tunisia and Kenya.
The Australian government has brought forward the nation's IPv6 cut-off date by several years, mandating a complete government transition by the end of 2012. However, our major ISPs have been quiet, though they are, no doubt, pedalling furiously behind the scenes to catch up.
Telstra started trialling IPv6 a year or two ago. It estimates the transition will cost it tens of millions of dollars and take at least five years.
Moving forward
So what to do? Consumers can only worry about their own points of connection. The rest is beyond their control.
Anyone purchasing new hardware or software that has any sort of internet connection should ask if it is v6-capable, because most modems and routers on the market are not and those devices that are can be costly.
The same philosophy applies to new computers, though by now, they should all be running v6-compatible operating systems.
Similarly, anyone signing up for a new internet or mobile phone plan, particularly the two-year plans, should ask the same questions.
Is your new hardware (back-end and the devices being deployed in your home or business) IPv6-capable and, if not, will it be upgraded — and at what cost to you?
Of course, the biggest question is whether we have learnt from the IPv6 debacle or if we will be having this same discussion 20 years from now.
"Never say never on the internet," Mr Biber says.
"But the number of IPv6 addresses is so astronomical and unfathomably huge, they will last us a long, long time.
"I'm confident that the basic IPv6 protocol will still be around in 20 years, though I think the internet will be quite different."
That's what they said 20 years ago.
Time (and IP addresses) running out
IP ADDRESSES are doled out by five bodies globally: APNIC, which disseminates addresses throughout Asia and the Pacific region; AfriNIC (Africa); ARIN (North America); RIPE NCC (Europe, the Middle East and parts of central Asia); and LACNIC (South America and the Caribbean).
Each body has its own pool of millions of IP addresses but the general global pool, the one that is running out, had only
204 million left as this report was written. With existing consumption rates, the global pool will be exhausted by about the middle of next year. However, the individual regional pools will likely have enough IP addresses to last until late January 2012, after which there will be no more IPv4 spots available.
What happened to IPv5?
DARPA (the Defense Advanced Research Projects Agency), an agency of the US Department of Defense, implemented IPv5 in the 1980s for a streaming internet protocol but it was a research experiment that is now redundant.