The next version of the ubiquitous Internet Protocol (IP) is coming, as sure as Y2K followed the 1990s bubble.
In the end, Y2K looked like a non-event, and it is widely dismissed today, but that is a misleading analysis. The world as we knew it did end; most just didn't notice. Much of the reason was that thousands of techies and out-of-work Cobol programmers had been employed to make sure the most critical problems caused by two-digit dates had been addressed, even if only clumsily. Most of the rest of us just dodged the problem by upgrading all our kit.
The collapse of dot com mania masked the real reason for the IT spending drought of the early part of this decade. Much of the downturn was caused by the synchronisation of enforced software and hardware upgrades, thanks to Y2K.
The next version of IP is confusingly known as version six. No, for all practical purposes, there wasn't a version five and version four was the first version.
The switch to IPv6 is likely to be another much-hyped, yet underwhelming, important, yet ignored, event. It will be, once again, the end of the world as we know it, and once again, the world's end will fail to impress.
The cause of the problem that IPv6 addresses is painfully familiar. Back in the day, it was thought that four bytes, providing four billion or so unique addresses, would be more than enough address space for the small research network we know today as the internet.
We're likely to be watching football when the apocalypse comes.
Ivo Vegter, freelance journalist and columnist
The problem has been postponed by making networks responsible for their own internal addressing (known as network address translation) - akin to assuming that a two-digit year should be prefixed by 19 if it is above 30, and by 20 if it is 29 or less.
IPv6, by contrast, will provide more addresses than we'll ever need. No, really: there will be 3.4 x 1038 unique IP addresses, or 50 000 000 000 000 000 000 000 000 000 for every single human on the planet. I'll eat my hat the day we run out again. Y10K - the need for 5-digit years - is likely to come sooner, by which time I'll surely have lost my hat.
As mobile technology and consumer electronics explodes, every product gets its own unique RFID tag, and every electrical device gets a chip and network location, trying to make today's kludge work forever is simply impossible.
There are other benefits to IPv6, most notably network-layer security, which makes encryption a low-level function rather than the poorly-used and optional application-level add-on that it is today.
The fact is that the world - ironically with China in the lead - will switch to IPv6 sooner rather than later. The US government has mandated it, Europe is moving ahead, and telcos are setting it up in their core networks. The South African address space has been allocated IPv6 address blocks, and the domain name administrator accepts IPv6 registrations already.
The problem is that one day, a switchover of all networks will be required. This will break a lot of stuff. That day isn't some distant future; it is likely to be around the middle of 2010, when IANA, the central body that issues internet addresses runs out of new IP numbers to issue. (In fact, we're likely to be watching football when the apocalypse comes.)
The industry has worked hard to make the IPv6 transition problem invisible to its customers. As an industry, nobody wants a repeat of the ill-understood hype around Y2K, since that permanently tarnished the industry's good name. In the end, however, it will affect everyone.
There will be transition mechanisms, that permit existing IPv4 systems to co-exist with the new IPv6 internet. Some will work better than others, and some (like tunneling IPv6 packets across IPv4 links) are downright ugly. There will be pushback, as incompatibilities and the increased overhead of handling bigger packets with bigger addresses prompt people to disable IPv6 while IPv4 still works just fine.
As doomsday draws near, however, expect a lot of upgrade cycles to be driven by the same “while we're at it” logic that drove Y2K upgrades. Expect a lot of people to make a lot of money solving urgent hardware and software incompatibility problems, and coining it from a great number of imaginary problems “while they're at it”.
And expect to see infamous names resurface; names you thought you'd seen the back of after the Y2K debacle. Anyone remember Peter de Jager and Karl Feilder? Come gentlemen, encore please.
Ivo Vegter is a columnist, freelance journalist. He blogs at http://ivo.co.za/.
Share