Analogies between the Y2K Millennium Bug and the need to switch the internet to IPv6 aren't necessarily helpful, we heard at last night's panel session of @tref and Timico's Bring On IPv6 party at the London Transport Museum last night.
|Conducting the IPv6 party, |
Timico's Trefor Davies
|Gary Feldman singing|
The Day The Routers Died
|Ed Vaizey (Minister for internet)|
|Simon McCalla, Nominet|
|The last IP address block|
|Net pioneer prof Peter Kirstein|
|Abigail Harrison wasn't gurning!|
IP - Internet Protocol - makes the internet happen, but we've run out of IP addresses and it needs to switch to a new address system.
The purpose of the party was to raise awareness, and the presence of government Minister with responsibility for the internet Ed Vaizey must have gone a long way to help.
One of the reasons Y2K analogies aren't helpful is that there's no obvious epiphany looming, contrasting with Y2K, which was caused by the impending deadline of 31st December 1999.
In some respects we've already passed the deadline. IP addresses are allocated by the Internet Assigned Numbers Authority (IANA) to regions from a central pool. Allocations are made in blocks - known as subnets - and on 3rd February 2011 the last block was allocated.
Soon, maybe as early as April, the first Regional Internet Registry (RIR) - organisations responsible for allocating IP addresses to ISPs - will run out of IP addresses, meaning no new IP addresses for ISPs.
But internet users worldwide won't notice a problem until their ISP is unable to offer them an IP address for their connection, meaning the user will be forced to share an IP address with many other subscribers using a technology called Network Address Translation (NAT).
Some subscribers using a NAT ISP may not notice a difference. In some countries, especially in Africa, standard services have always used NAT, with customers paying a premium if they wanted their own IP address.
But many customers will notice. Online gamers, users of some internet telephony services, businesses, publishers will probably be the first affected.
And of course, no new IP addresses effectively means no new ISPs in a region once the RIR responsible has exhausted its pool, dampening competition in the market and in all likelihood leading to higher prices.
Analogies with the millennium don't work. Critical systems will not fall over, as they will always - and I mean always - find a way of getting the IP addresses needed. There is no well-defined point at which the ordinary user will notice a problem, and the emergence of problems will be localised by region.
In addition, the switch-over is far harder than making computers Y2K compliant, as it requires co-ordination and co-operation between ISPs, web hosts, web site operators, equipment manufacturers and subscribers to fix a problem most people don't even know exists.
The new system must work alongside the existing system, applies to every device that connects to the internet - security systems, set-top boxes, mobile phones, printers... - plus there's many more computers in existence now than in 2000.
The IPv6 problem won't solve itself, and we can't expect any major impetus from national governments until problems such as competition in the market and ISP pricing get severe enough to force action. And at that point, it will still take significant time - several years - to complete the transition.
There is however 2 similarities with the Millennium Bug. 1: The problem is real, and IPv6 switch-over must happen, and 2: many software engineers, especially those with experience developing real-time embedded systems ;-), might find an abundance of short-term job opportunities!
On the plus side, engineers love solving problems. With a momentum of support from the goodwill of the tech community driving the switch-over, the problem might just get solved - before too many people notice.