Q&A: Why switch to IPv6?

Network cable connector

Fewer than 10 per cent of IPv4 addresses are still available, and they're set to run out by 2012, according to the Number Resource Organisation (NRO).

IPv4 is the internet protocol numbering system currently in use, but it's set to run out of numbers within in the next two years. There's a new system in place - IPv6 - but businesses haven't been switching over quickly enough.

We spoke to Axel Pawlik, chairman of the NRO, to find out more.

What's the issue with IPv4?

In about two years time we will be running out of IPv4 addresses. The fact that the four billion addresses we have there are quite finite was known from the start, when the internet started to happen in the late 70s and early 80s.

It was not really expected that it would be such a smashing success, as it currently is. It was thought to be a tool for scientists and computer geeks. Now, obviously, that has changed.

We currently see on average two IPv4 numbers per head in the more developed world. Now that there are so many billion people in the world, if you were to interpolate that up, there are just not enough IPv4 addresses for everybody.

And it's not just per head. Every person has pockets full of gadgets, and we have our fridge at home and our TV, and so many other computer-like things, it's not going to last.

How long do we have until they run out?

Currently we think the IPv4 addresses we will give out until sometime in 2012. That's based, of course, on a bit of guesswork and interpolation on what we knew about previous allocation rates.

That might change. We might get one or two bigger requests that we currently don't foresee, and then we might run out a little bit earlier.

So what's IPv6? How does that solve the problem?

In the 90s, the engineers thought about a way around this, and they called it IPv6. That's version six of the protocol. It's basically the same function, just more addresses. Many many more addresses

The idea was many, many years ago that everyone would pick up this new version and would run that, and if everybody has it running, then we would be able to dump the IPv4 stuff because nobody would need it anymore.

But that was based on the assumption that people would actually pick it up quickly. That hasn't happened, and now we are coming to a crunch the growth rate is quite high, and now we must adopt IPv6

If we don't act relatively rapidly, over the next couple of months, start doing something in earnest, what will happen is that the internet might either not grow as fast as it used to - which is not nice as you want to hook up new people all the time - or there is the quite real danger of added complexity and added costs. And, in the end, also [the] added risk of failure.

Are there other reasons why we should be switching to IPv6, aside from running out of numbers?

That is the main reason. There is no great added functionality to IPv6, apart from just having more addresses. We have to if we run out of bread, we should eat rice.

But some people say IPv6 offers more security. Is that not true?

It's not true, really. Basically, all the security you can get with IPv6, you can now have with IPv4. It's not exciting to go to IPv6. There are no new games, no killer application that runs with IPv6. It's infrastructure, it's good housekeeping. Proofing for the future.

Is anyone leading the way with this? Are businesses or government doing better?

Nobody really is leading. That's the point.

The problem here that we all know exists is that there's no immediate business case to it. Now if you have to change something, that will cost either equipment or software or training.

There is some cost involved, and there's no customers actively asking for IPv6 because there is no added functionality, there is no push really. We only know that in two years time, there will be a crunch if you haven't prepared for it

We quite a number of questions over the last couple of years from governments: "oh we see this, lots of our economy is based on the internet, it's critical infrastructure, what can we do to make this happen?"

Well, not really that much. Not really that much. If you are in Western Europe, you can lead by example, of course go ahead and talk about it and make a campaign about it, but basically I think governments should lead by example and have their websites and their services available on the internet on IPv6.

What costs are involved? It's just switching IP addresses, really.

That really is it. If we look at where we are now, and IPv6 is no big news anymore, basically all the available operating systems, software for your PC or for your Macs are able to run IPv6. It's built in, all you have to do is switch it on sometimes you don't even have to switch it on, it's switched on by default

There is certainly an amount of training for the engineers that would have to run the network, if you're a network provider. But again, it's not a functional difference. It's just longer addresses, you have to get used to them, but really that's all.

IPv4 and IPv6 don't run side by side then?

They do exactly that, they run side by side. But they don't talk to each other. Which is a silly idea, somebody introduced it 15 years ago or so, when IPv6 was designed.

The engineers were basing their design on the idea that there was so much time, that everyone could have IPv6 in parallel with IPv4, and then when everything was united, then we could dump IPv4. That won't happen that way anymore.

In that sense, it was a silly decision to say they are not compatible. You could have made IPv6 with variable length addresses, where IPv4 would be a compatible subset of IPv6 addresses. That would have been brilliant. Then we wouldn't have this problem now.

Read on to find out when to switch to IPv6.