IPv4 is obsolete.
Maybe it seems a little rash to say that. Maybe you think I’m nuts. After all, around 82.6% of IP traffic in the United States is still IPv4, as is the great majority of internal traffic of most organizations, be they enterprise, academic, or government. Nevertheless, IPv4 is obsolete.
The Internet Assigned Numbers Authority and all Regional Internet Registries except AfriNIC have declared their IPv4 address pools depleted. You can’t get any more IPv4 address assignments from them. You might be able to still get a limited assignment from your Local Internet Registry, but that’s likely to be from a small, protected pool that will be gone soon. You simply can’t get a meaningfully large block of IPv4 addresses anymore.
Well, that’s not entirely true. After all, you can still get a Model T if you have enough money. But you won’t be getting it from Ford. In the case of IPv4, address brokers have arisen over the past few years that can help you find and purchase IPv4 address blocks from other organizations. Count on paying through the nose for them.
The IETF is already beginning to plan for the decommissioning of IPv4 through the Sunsetting IPv4 (sunset4) Working Group, and has produced an Internet Draft that proposes moving IPv4 to historic status.
There are also discussions within the IETF to standardize IPv6. That means that when we talk about IP generically, we mean IPv6. Geoff Huston, Chief Scientist at APNIC, has written a couple of thoughtful blogs on declaring IPv4 historic and on declaring IPv6 an Internet Standard. He’s skeptical of both, but the fact that these discussions are happening at all should give you pause when thinking about the future of your own network addressing.
You might be inclined to say that as an enterprise, what the IETF declares is meaningless to you. You have enough IPv4 space left to sustain network growth for decades.
But is that true? Ask yourself:
- How many customers are accessing your data centers or online services from homes or small offices?
- Do you serve customers in Asia or Europe?
- Do you plan to deploy new branch offices overseas?
- Do you have employees telecommuting or accessing the network while traveling?
- Have you considered how virtualization and containerization will impact your existing address supply?
- Have you considered how the Internet of Things (IOT) might transform your business? And the number of addresses all those “things” will require?
And by the way. Those questions about Asian and European customers and office locations are significant now only because APNIC, RIPE NCC, and LACNIC declared their IPv4 pools depleted earlier than ARIN did (in 2011, 2012, and 2014, respectively). But ARIN is done now too, so consideration of customer and office locations soon means IPv6 no matter what part of the world.
I’ve been deeply involved in advocating IPv6 since the late 1990s. Back then and up until a few years ago, the primary audience when “preaching IPv6” was service and content providers, because they would be – and have been – the first businesses impacted by IPv4 address depletion. The objections evolved through the years, but the most commonly heard were:
- We believe it will be decades before IPv4 is exhausted.
- We will support IPv6 when we see a demand for it (the infamous “chicken and egg” dilemma).
- We understand that we will need it eventually, but we have more pressing priorities. We’ll get there eventually.
- Implementing IPv6 is just a matter of toggling a feature, isn’t it?
- Implementing IPv6 is expensive and complicated. We need a more compelling business plan.
The contradiction of the last two objections is particularly interesting. It indicates an impulse to avoid an oncoming problem more than a willingness to face it. Head in the sand. Nevertheless, virtually all service and content providers have overcome their objections, recognize the inevitability of IPv6, and have either deployed IPv6 or are well on their way to doing so.
Enterprises, the last IT industry group impacted by IPv4 depletion, are beginning to ask the same questions service providers were asking ten years ago. And often, they’re voicing the same objections. But enterprises are in a much better position to face the transition from IPv4 to IPv6 than service providers were:
- There is no speculation about when IPv4 pools will run dry. They’re gone.
- Vendors with no IPv6 support are much less of a roadblock to deployment now. In fact a vendor that does not support IPv6 in their product, or requires special IPv6 licensing, should be seen as having a flawed IP implementation.
- Because of the lessons service and content providers have learned, there is now a far deeper knowledge base supporting IPv6 deployment.
IPv6 is inevitable. Running dual-stack networks is complex, expensive, and risky, and brings us to the “tipping point” where we will begin actively pushing IPv4 out of our networks. The sooner you deploy IPv6, the sooner IPv4 can ride off into the sunset.
Fishtech has deep experience with IPv6, and can help you with systems evaluation, deployment planning, address design, and many other aspects of IPv6. Contact us to discuss IPv6 in your network.