IPv6 – are you ready? Pt 1


SAGE-AU is a guest blogger.

SAGE-AU is a not-for-profit professional organisation that promotes the development of the system administration profession.

Part one: the problem

The Internet as you’ve come to know it is about to change. It’s not a bad change, and it won’t be a dramatic change, but it’s a change that if you work in this industry, you’ll need to know about – preferably now. Yesterday, actually.

Specifically, current estimates are that the current Internet Protocol – known by network engineers as IP version 4, or IPv4 – will run out of completely vacant, allocatable addresses in November 2011. This is, roughly speaking, the point at which it’ll be no longer possible for new customers to connect to the Internet in the way we’ve all become used to.

For starters, ISPs will likely have to start using various tricks to get around this; you may well expect to see some ISPs offering their cheaper plans’ customers a private address behind a network address translation (NAT) firewall, for example, to extend the life of their IPv4 address ranges, since there won’t be any more to acquire any other way. This is a definite “here be dragons” moment, particularly as many home users already use NAT as that’s how many home routers are configured, but ISPs will nonetheless need to find a way to make things work, and NAT is something they can implement relatively easily.

The biggest problem with NAT is that it breaks one of the fundamental ideas behind the Internet: any to any communications. With NAT, there’s no easy way for applications to talk directly to each other across the Internet; this can cause trouble for a host of commonly used tools such as various Voice over IP protocols, remote user assistance applications, and of course any “peer to peer” application. Many developers have worked around this issue, but the potential advent of two layers of NAT (one in the ISP, one in the home) has the potential to break a lot of functionality.

The root cause of the problem stems from the way IPv4 addresses are handled: in theory, IPv4 supports 4.2 billion unique devices on the Internet. Unfortunately, as well as being insufficient for the number of people on the planet (or even half of them, if you assume that half have a mobile phone and may want to browse the Internet on it), IPv4 makes many assumptions about how it is to be used that leads to wastage.

Given that it’s a thirty year old protocol this year , these assumptions are neither unexpected nor to be frowned upon. The end effect of this is that IPv4 can support a lot less than 4.2 billion devices, and we’re now running out of space for more. Think about your home; if you’re reading this article, chances are you have more than one connected device. More than one computer in your house? Well, you’re probably using NAT, but you’re definitely using more than one IP address of some form at home. Have a mobile phone, particularly a smart phone? There’s another IP address in use right there.

So here’s where we’re at:

1. The Internet as you know it is running out of addresses for new devices, and will almost certainly run out completely during this calendar year.

2. There are workarounds for this problem, but – as for workarounds of almost every flavour – they’re not necessarily palatable, or very good permanent solutions.

Of course, there’s a perfectly good technical fix for all of this, supported by every modern operating system and most mid to high end network equipment: IP version six, or IPv6. IPv6 sidesteps the main issues nicely, by providing about 3.4 × 1038 addresses – yes, just a few orders of magnitude more addresses than provided for under IPv4. As a side effect, this neatly does away with NAT and its foibles – there is, at least for the foreseeable future, no reason that every device connected to the Internet can’t be directly contactable.

And IPv6 is here, working, now: at least one major Australian ISP (SAGE-AU platinum sponsor, Internode) has comprehensive IPv6 connectivity on the Internet. Many major organisations, including of course Google, Facebook, and others already support IPv6 in some form.

Of course, there are good reasons that the transition hasn’t been happening with more gusto. First, IPv6 is not IPv4; there’s no real compatibility between the two. For now, running IPv6 when much of the world still uses IPv4 means having both enabled in your environment; fortunately, IPv6 makes this quite simple to do, provided you have the right equipment.

Equipment, of course, is the second issue: most ‘consumer’ grade ADSL equipment still, in 2011, does not support IPv6. This is changing – with a firmware upgrade Billion’s 7800NL modem supports IPv6, for example, as does the FRITZ!Box – but for the most part, there’s precious little within the price range most consumers will buy in.

This in turn seems to be related to most ISPs’ reluctance to engage in widespread IPv6 testing: without consumer equipment, or consumer interest for that matter, most ISPs are at best testing the water with their business customers (and even there, take-up appears to be at best weak). In any case, if you’re involved in the Internet industry, now is the time to be looking at your IPv6 deployment plan. More on how to do that next column.

By Iain Robertson

Iain Robertson is a senior network engineer with a large research and education provider in Queensland. He has over a decade experience working in various sectors of the IT industry, and is an active member of SAGE-AU - Australia's peak professional organisation for system administrators.

Tags: ipv6, SAGU-AU

Show Comments