Experimental feature

Listen to this article

Experimental feature

Here’s why your can of cola, indeed any of your possessions, could soon have their own internet addresses: a new protocol, IPv6, which will make this possible is on its way.

Last month, the latest in a grand series of conferences designed to smooth the path of IPv6 was held in Taiwan. Yet outside academia and the US and Japanese governments, interest in IPv6 is virtually nil and there are almost no commercial deployments.

It’s a poor reward for the internet technical specialists who have been working on IPv6 for more than 10 years and who believe that without its adoption, development of the internet will slither to a standstill.

Internet protocols are the sets of rules that define how different computer networks talk to each other.

The version that you, me and the rest of the world are using is version 4, now almost 20 years old and stable and reliable. Version 5 was experimental and never adopted, so the digit was skipped to avoid confusion.

Why is there a need for a version 6? Chiefly because the internet is being used for all manner of communication – telephony, video, fridges that call for replenishment when empty, and so on – which the architects of the original internet had not allowed for.

Every piece of intelligent electronics attached to the internet requires a unique address and version 4 will not be able to provide enough of them.

It uses addresses which are 32 data bits in length: enough for only about 4.2bn unique addresses.

IPv6 addresses, however, are 128 data bits in length giving, according to one calculation, six times 10 to the power of 23 addresses for every square metre of the Earth’s surface, about 340 trillion in total.

As consultant David Morton put it: “This capacity doesn’t just allow an IP address per toaster, or even an IP address per slice of bread in a multi-slice toaster, it is far greater than that.”

There are technical reasons to do with the ease of directing packets of data through the internet to their correct destinations that make this level of redundancy a good thing.

But there is also the point that the effort and cost involved in migration to IPv6 is going to be huge.

Neil Rickard, a senior researcher with the Gartner Group consultancy, says that the billions expended on the Millennium “bug” was trivial by comparison.

So nobody wants to do it twice. At present, however, very few people seemingly want to do it even once.

There has been some activity in Japan where tax breaks make it cheaper to build version 6 networks than the version 4 variety.

And the US government is mandating the Department of Defense and encouraging other agencies to convert to IPv6 by 2008.

The problem is that for most commercial organisations, the new protocol brings few benefits to justify the cost.

There are some improvements in security, in network management and in quality of service but not enough to make the argument compelling.

Moreover, most companies have no problems with the availability of addresses. They use network address translation (NAT) software which assigns a single IP address to the company, behind which it allocates private addresses to its equipment. So the commercial case is yet to be made.

Mr Rickard worries that priority for countries such as China and India and industries such as the third generation mobile phone business, which need lots of new addresses, will get lost in a “me-too” style rush to keep up with the US.

In any event, however, it seems that version 4 and version 6 will have to co-exist for years. Fortunately, communication between the two protocols is comparatively straightforward.

By that date there should be no need to use the drastic measures adopted by Vinton Cerf, the “father of the internet”, to persuade recalcitrant network users to switch to the new-fangled protocol.

The internet originated in a small military network called arpanet. It used Network Control Protocol (NCP) to which many of its users were devoted.

Mr Cerf and his colleagues devised a new internet protocol known as TCP/IP in 1978 and decided in 1980 that the whole network would use the new protocol. At the time, the two were running in parallel.

One day, Mr Cerf and his colleagues turned off NCP. As he recalled: “This caused a lot of hubbub unless you happened to be running TCP/IP”.

But NCP users weren’t persuaded that MrCerf was serious until he switched off the protocol first for two days and then forever.

Since then IP has grown into the de facto international networking standard and IPv6 will maintain that status.

Get alerts on News when a new story is published

Copyright The Financial Times Limited 2019. All rights reserved.

Comments have not been enabled for this article.

Follow the topics in this article