What has happened to IPv6?
The logic of the move over to the new Internet Protocol version 6 (IPv6) is undeniably good, and it goes like this – while IPv4 wasn’t a bad solution, it’s now old and tired, and in almost every area you can find things that would be done better if only the protocol had been better designed back in the early days of the internet. And by early, I don’t mean the first time you clicked on a Google link, I mean going right back to the pre-history of the internet some 30 years ago. The trouble is that most people don’t really care about the problems of IPv4. You might be surprised by this, but the truth is that none of us chose IPv4, or any other version: we ended up with TCP/IP as the result of a ferocious but unglamorous competition with various other protocols – those with long enough memories may remember that protocol we used on Windows for Workgroups, for example. I don’t want to give the impression that TCP/IP was just the best of a bad bunch, though, because it’s a remarkable piece of work and the single most enabling technology that allowed the internet to come alive. It really has done jolly well by us, but it does have some limitations.
For example, we’re constantly being told that we’re running out of IP addresses and that there will soon be a huge crunch. This is true, but only up to a point Lord Copper. A number of huge companies have ended up owning mind-bogglingly large IP address spaces (called Class A addresses in the old name-speak), and they have no real intention of giving these up. In the old days, we used to be given real live public IP addresses for each computer, but for most of us little people this state of affairs has nowadays been subsumed behind a single public-facing address using the technology called NAT (Network Address Translation). IPv6 has been designed to do away with all such limitations: it supports such a vast address space that a hundred thousand IP addresses can dance on the head of a pin or some such sillyness. All of which is a good idea, together with new built-in capabilities for quality-of-service management and so forth.
So where is it? Every decent OS now supports IPv6 and has had drivers for a number of years now, but the more you look, the more it isn’t there. There are a few companies that have implemented IPv6 in a big way – my sources tell me Microsoft has for example, as has Google, but neither of those is really a surprise, since they both need to run huge datacenters. Microsoft also writes operating systems, so running IPv6 is more or less imperative. But out in the public space there’s no sign of it. I’ve yet to find a single organisation that’s using it. Sure, there are little puddles in academia, or in some specific datacenter farms, but in the mainstream there’s nothing.
So when will IPv6 arrive for real? Who knows? Some ISPs tell me that they might be seeing IPv6 support on their wire some time around the end of next year, but there’s no great rush for it. If your business thinks differently (and I’m not looking for the telcos here), please let me know the when and the why.
A friend of a friend wanted a simple but quite sensible facility – the ability to run all their big print jobs during the evening when the office was locked up and closed. During the day, the single big colour laser printer is used to print out paperwork that’s needed right now, so having to sit around and wait while someone prints a 500-page document is tiresome and gets in the way of everybody’s workflow. This seemed like a perfectly logical and straightforward requirement, and he’d approached it by applying a little knowledge in entirely the wrong way.