Seriously, you need to think about IPv6

Although everyone seems to ignore the inevitable, IPv6 will become a reality soon, so plan for it now

I've written several times about the inexorable decline in available IPv4 space and why that means we're all going to have to get our heads out of the sand and learn IPv6. If you read that particular piece -- written more than two years ago -- you'll recall that I opined that it'd only be a couple years before we'd all be forced to learn IPv6. Of course, that hasn't come to pass. ARIN (American Registry for Internet Numbers) still has more than two /8 IPv4 blocks (each with more than 16.7 million addresses) of space available to be allocated, and there's undoubtedly still a lot of slack space in the allocations that have already been made to ISPs. So things have not reached the crisis level that I imagined they might two years ago.

That said, North American ISPs are starting to tighten their belts with IPv4 allocations as they see the later phases of ARIN's IPv4 Countdown Plan near. As that happens, it will become increasingly difficult for ISPs get IP space from ARIN, and so it will be more difficult for downstream organizations to get space from the ISPs. So even though we might not actually ever be "out" of IPv4 address space, its scarcity will have far-reaching consequences.

One recent example I ran across was Northeastern U.S. ILEC FairPoint's use of /31 network blocks for handoffs to business DSL customers. Their use was obviously an attempt to reduce the consumption of address space, because these blocks are incompatible with a large swath of customer premise equipment (typically, only real routers know how to use these tiny two address subnets, which lack a network and broadcast address found in the more typical /30 block). Make no mistake: The early effects of aggressive address space conservation are already being felt and will only get worse.

So although it might have been alarmist to suggest that 2013 is the year when we will run out (realistically, we're looking at April 2014), it is high time to get a plan together for how you will cope when you can no longer get IPv4 space. That's not necessarily to say every enterprise must immediately run out and deploy IPv6. Instead, simply having a plan put together now for when it inevitably does become a requirement makes eminent sense. After all, the outer limit of IPv4 availability is starting to creep within the scope of many of the budgets that enterprises are starting to assemble now, so there's no time like the present.

There are a few basic questions you'll want to consider as you plan your organization's migration to IPv6.

Do you have the skill set in house to make the shift to IPv6?

Above all else, IPv6 takes a different mindset to implement than IPv4 does. Many of the lessons for implementing a solid IPv4-based network still apply, but many other learned habits do not.

For example, NAT (network address translation) was initially used as a means to conserve IP address space, not as a means to implement security. Over time, NAT has became a sort of security crutch that separates internally addressed devices from the publicly addressed Internet. But with the mind-blowing amount of address space available with IPv6, NAT will largely become a thing of the past. As a network engineer, I can tell you that simply wrapping your head around the fact that every device on your corporate network will be addressed with a publicly routable IP address is more than a little disconcerting at first.

Critically, getting a good handle on how IPv6 should be implemented in an enterprise and what caveats and pitfalls there are is something to learn right now -- well before the planning for your inevitable migration to IPv6. It benefits no one to learn new skills right in the midst of planning a potentially months-long migration project.

Where will you get IPv6 transport?

The next thing to evaluate is your current ISP's willingness or ability to provide you native IPv6 transport. Although you can use IPv4-to-IPv6 tunnels to deliver IPv6 Internet access to anywhere, the best solution is to have a native IPv6 handoff. These days, most ISPs worth their salt will offer you native IPv6 connectivity, but some still do not.

Because IPv6 is designed to operate in parallel with, but separately from, IPv4, you may be able to implement IPv6 in a provisional, testing capacity right now and make the process of learning how it can be used that much easier for when it's required.

Will your network devices and software support IPv6?

Broadly speaking, there are very few network devices in use on a modern enterprise network that do not support IPv6. However, basic support for IPv6 addressing in network hardware and operating systems isn't the only concern. Software, too, can be a problem -- a lot of enterprise software is written using older APIs that are not IP-version-agnostic. The impact that this may have could be relatively minor -- when, for example, the application needs to operate over only an internal network where IPv4 address space conservation typically isn't a problem. However, it could be far more severe if the application needs to serve requests from the open Internet or reference such addresses internally.

Lessons from Y2K

This IPv6 situation we find ourselves in today is not all that unlike the Y2K scare that happened in the leadup to the year 2000. Many people look back on that and laugh, as not much of anything really happened when Y2K finally did roll around. However, those who were working in enterprises -- particularly larger ones with substantial internally developed code bases -- will remember the massive amount of effort that went into ensuring that nothing bad happened and that the lights stayed on. The Y2K non-event required a lot of effort, effort that seems lacking in the IPv6 transition.

The exhaustion of IPv4 is different from the Y2K scare in that it will not take place all at once. But there's also no firm date for when it will reach criticality, meaning there's no clear deadline for people to aim for. And without real deadlines, things slip -- often too far. Enterprises knew for many years exactly when midnight on January 1, 2000 would roll around. But no one can tell you with any certainty when Tier 1 ISPs will run out of IPv4 address space to give to customers -- or when you'll simply become too small of a fish for them to consider doling out their remaining space to. (That could be as soon as next year.)

So now is the time to do the learning and planning, if you haven't already. Don't be caught flat-footed when IPv6 finally becomes something you can't ignore anymore.

This article, "Seriously, you need to think about IPv6," originally appeared at Read more of Matt Prigge's Information Overload blog and follow the latest developments in storage at For the latest business technology news, follow on Twitter.


Copyright © 2013 IDG Communications, Inc.