Prisoner 'No.6' (Patrick McGoohan)

IP Version 6 migration

IT strategists should be planning for their organisations' transition to the impending introduction of IP Version 6.

Imagine that you are the IT administrator charged with IPv6 roll-out at your organisation. IPv6 is the new iteration of Internet Protocol designed to replace IPv4, the protocol on which the Internet is based. You've known about IPv6 for years, it's something that's been spoken about often you've even made plans for how it is going to be adopted but something's always got in the way; and during the latest economic downturn, only projects that can show an guaranteed return have been approved. And now time is running out fast. IPv6 panic is starting to creep in. You may even wonder if you are over-reacting; and you may worry that you are not.

The stark fact is that any IT professional who is not planning for IPv6 migration is unworthy of the job title. Organisations are being exhorted to rethink their plans for the Internet because IPv4 addresses have a limited lifespan. When IPv4 was first ratified, when the Internet was still at a relatively early stage of development: little thought was given to the fact that the 4,294,967,296 addresses offered by the protocol would eventually become inadequate. This was a time before the mass-adoption of the Internet for commercial purposes, before the boom in mobile devices, and before the rise of India and China as demanding commercial powers developments not anticipated 30 years ago, but which have led to many of the issues the Internet community now faces. In addition to Internet-connected computers and smartphone, many household appliance could son be deemed deserving of a dedicated IP address at times; and the predicted rise of smart metering computerised utility meters that use the Internet infrastructure to send and receive usage data will also add massively to IP address demand.

These factors and more besides have placed a tremendous demand on the number and capability of the available IPv4 address; and unless something is done about it the Web would enter an era of prolonged stasis, where the Internet becomes a business inhibitor, rather than a business enabler.

We are a ways off zero hour yet, but on 19 January 2010, the Number Resource Organisation (NRO), the official representative of the five Regional Internet Registries, announced that fewer than 10 per cent of all IPv4 addresses remain. The NRO took the opportunity to urge all interested parties ISPs, vendors, and users to accelerate IPv6 adoption before they ran out.

The dwindling number of IPv4 addresses has not been ignored. There have been warnings about address shortages for years. In 2005 Cisco as a leading networking vendor not without a vested interest in the matter predicted that IPv4 addresses would dry up by 2010. This has proved to be a little wide of the mark, but no-one's sure by how much.


According to the University of Southampton IPv6 project (the university was one of the pacesetters in the technology), there are just 11 servers in the UK that are running IPv6. And there is little sign that the UK's Internet Service Providers 'ISPs' are set to launch trials of the technology, although one, Andrews & Arnold has been an enthusiastic early adopter of the technology. JANET the UK's renowned Joint Academic Network is also IPv6 ready, and has been involved in IPv6 deployment since 2003. In the US, however, cable TV operator and ISP Comcast has just launched its initial IPv6 trials the first major player to do so and as consumers will be part of the trial, it seems that there may finally be some mass awareness of the problem.

Not that public awareness has necessarily been to blame for under-adoption. According to some commentators, the long lead-in period has led to the state of inaction. For Axel Pawlik, chief executive of RIPE NCC, the European Regional Internet Registry (RIR), the issue has not been with the network providers or the vendors, it has been the demand from the users, both business and consumers, that has introduced the delay.

'[Moving to IPv6 has] been talked about for years, but that's precisely the worst thing that could have happened,' says Pawlik. 'What that meant was that there was a degree of panic-mongering about the imminent shortage of IPv4 addresses, and then when that didn't happen, it was hard to get people to take the issue seriously.'

Pawlik, who also chairs the number resource organisation within the RIRs, avers that awareness has been growing as organisations have realised, gradually and disconcertedly, the hard truth that there is no alternative to IPv6; not that he believes that the Internet protocol's sixth incarnation is actually the best of solutions. 'In our opinion, it is still not fast enough,' he says, 'but when you think that in the beginning of the process, we weren't seeing any awareness of the protocol at all, then that is some progress.'

Impact of IPv6

One common misapprehension about IPv6 is that its influence begins and ends with network functionality. In fact, IPv6 is designed to fulfil a more extensive role in the connective process. Knowledge of the protocols has to go much further: organisations with Websites (that's pretty much everyone) should ensure that they support IPv6 too. Understanding of the reasons for this necessity can be sketchy, even among well-informed information technologists. What's more, the ignorance quotient is not necessarily highest among the end-user sector; content- and service providers need to upgrade their knowledge. John Curran, president and CEO of the American Registry for Internet Numbers (ARIN), has warned that 'when we get to the end of 2011, we are going to have a lot of people connecting over IPv6, and that does not bode well for the content providers who don't support IPv6.'

RIPE NCC's Pawlik agrees. 'The techies are pretty much aware of IPv6 now, we've having to reach further,' he says. 'We have to reach out to the business people to raise awareness with them so that it is the CEO who is asking the techies do we have to do anything?. That's the next stage.'

But the prevarication is reaching its end now, he warns; absolute deadlines are looming and looming large. There are two dates we all need to bear in mind, Pawlik says. Sometime in mid-2011, the Internet Assigned Numbers Authority (IANA) an adjunct to ICANN - will allocate the last IPv4 number. The RIRs will continue to work with those numbers, but some time around July 2012 Europe will run out, although Pawlik expects Africa will continue to allocate 'much later'.

Leo Vegoda is number resources manager at IANA, agrees that most ISPs are fully ware of what needs to done. 'As to whether organisations without a deployment mandate are heeding the calls for IPv6 adoption well, I think it depends on their particular circumstances. The full allocation of the IPv4 address space is first going to hit the networks that want to continue growing the hardest.' He sees that those companies will probably not be relying on IPv6, but will (initially) depend heavily on protocol translators run by ISPs.

IPv4-to-IPv6 translation 

The drive to develop IPv4-to-IPv6 protocol translators is going to be interim state that many organisations will depend on while the world waits to move to IPv6. There have been several moves to address the problem of IPv4 to IPv6 translation. In the Comcast trial, the company is examining three alternative methodologies. The first phase is going to use a technology from a French ISP called 6rd, which tunnels IPv6 traffic over IPv4 addresses. The second phase will use both native IPv4 and IPv6 traffic running side-by-side. This is an approach called Dual Stack, and is a technology that has been modified for the third part of the trial which Comcast's own technology, called 'Dual Stack Lite', which uses network-address translation to share a single IPv4 address among many customers.

Interesting about the Comcast trials is the fact that it appears to have taken the same tack as Pawlik suggests, and gone about reaching beyond the techie community; and the company is looking to attract a wide range of customers to the trial. In the public blog announcing the trial, Jonathan Livingood, Comcast's executive director of Broadband Systems says.

'The 'Internet society' does not believe that stakeholders have the luxury of taking a wait-and-see approach, or in indulging in circular 'who goes first' discussions,' he says. 'The Internet society appreciates that many players such as network operators, hardware manufacturers, handset manufacturers have made great strides in engaging on this important issue, but, as most would agree, much more needs to be done by all stakeholders.'

Comcast has, therefore, invited any of its customers to participate in the trials, which will be free of charge. The company is catching-up with another US company, co-location and hosting specialist Hurricane Electric which launched its Tunnel Broker: this allows users to reach the IPv6 Internet by tunneling over existing IPv4 connections, providing the user's host is correctly IPv6-enabled.

But approaches such as Comcast's and Hurricane Electric's may be masking the problem. According to Qing Li, principal engineer of Blue Coat Systems, these do not help people within enterprises who are trying to deal with applications.

'Many of the IPv4 translation initiatives focus on the network and tunneling, which doesn't necessarily mean that the applications are delivered effectively,' he admits. He points out there was also a degree of complexity in the approach. 'Comcast requires you to upgrade the client machines, and upgrade the server too, so that is extra work it is not very efficient,' he says.

Blue Coat's solution to this problem is to focus on application delivery. Its IPv6 secure Web gateway solution acts as an intermediary between requests and content, making the retrieval of applications, services and data whether IPv4 or IPv6 - environment transparent to users. Blue Coat claims that this means that users can work with minimal disruption. Qing Li sees it as a way to deal with the fact that many people are still not fully acquainted with IPv6. Like many organisations, he sees that interest is picking up.

Axel Pawlik is also dismissive of the move towards protocol translation. 'Of course, there are these initiatives that offer some sort of protocol translation, but I don't see them as being helpful. They are just technical crutches.'

Surprisingly, Pawlik's fear is not that such approaches will not work, but that they'll work too well, and slow down the migration to IPv6. 'It will work too well for a time, that is our biggest fear, but that will not last forever. It will just get more and more complex to keep the system running,' he counsels. 'The simple solution is just to move to IPv6.'

Pawlik adds: 'In a few years time, there will be people who will only have IPv6 - and what are the others going to do then?'

Sign up to the E&T News e-mail to get great stories like this delivered to your inbox every day.

Recent articles