Since around 2000, analysts have pointed to the accelerating "everything over IP" trend and warned that we had better get ready for IPv6. So why, 11 years later, are we are still nagging many networking vendors to please, just add IPv6 support to their products already?

Since 2006 we've known that 2011 would be the year IPv4 addresses would be depleted, yet so many of us were caught flat-footed by the announcement that--surprise!--the very last IPv4 addresses had been handed out to the regional registries.

How did this happen? It's really not all that surprising. We tend to prioritize our network upgrades around the most urgent needs, and IPv6 has always seemed like something that could be put off just a year or two longer while we focus our budgets and engineering resources on that new MPLS core or those new data center switches or whatever else appears to present a more tangible return on investment.

So does IPv6 finally have the requisite urgency to move at least close to the top of our priority lists? Yes, in the service provider arena, where the business case is straightforward: We are running out of an essential network resource (addresses) and must do something about it. But in most of the enterprises and content providers I work with, that urgency has not yet hit home. Internally, most enterprise networks can continue to grow using private IPv4 addresses behind IPv4-to-IPv6 Network Address Translation (NAT44), and externally accessible services do not grow quickly enough for IPv4 address depletion to present a near-term resource shortage.

Nevertheless, as I discussed in a recent InformationWeek Analytics report, externally accessible services are behind the business case for IPv6 in enterprise and service provider networks. Service providers, particularly broadband service providers, will soon be provisioning thousands of new home and small office customers on IPv6.

There are three alternatives for those IPv6 customers to reach online content and services:

-- The broadband service provider will provide a centralized NAT (Large-Scale NAT, or LSN) system that allows the user to reach IPv4 content using private IPv4 addresses.

-- The enterprise will stand up IPv6 proxies in front of their IPv4-only servers.

-- The services can be made accessible natively by IPv6.

Lab testing and some field experience have revealed that LSN and IPv6 proxies will adversely affect or completely break some applications. If your customers experience problems with your online services, they are not going to blame their service providers. They are going to blame you.

Therefore, the case for IPv6 in the enterprise and content provider arena is not the continuation of addressing resources, but the continuation of customer quality of experience. You can hope for the best with service provider LSNs (completely out of your control) or IPv6 proxies, or you can bypass those temporary workarounds completely by making your services natively accessible via IPv6.

I know what you're thinking--you're not feeling pressure from large groups of IPv6 users yet, so there is the temptation, as usual, to put off IPv6 deployment in your edge services for another year or two.

That would be a mistake. Every IPv6 project I have been involved with has been more complex than originally expected, and most of that complexity has stemmed from the need to identify and test applications in an IPv6 environment. Your IPv6 planning should begin by determining what applications, content, and services must be externally accessed. Are those services IPv6 ready? How do you certify them?

Who's responsible for modifying applications that don't behave well over IPv6? What are the criteria for determining whether to modify a noncompliant application or to replace it?

And don't forget your telecommuting employees, who must access internal services over VPNs. Will your systems be ready when their home broadband providers begin rolling out IPv6?

Sure, you can find plenty of reasons to put off IPv6 planning until next year, or to implement a workaround that appears to get you off the hook for the near term. But IPv6 is inevitable. Avoidance now will eventually cost you more money, in lost customers and lost productivity, than it will cost you to begin IPv6 implementation now. Trust me on this.

Learn more at the InformationWeek Analytics Live session at Interop Las Vegas on Thursday, May 12, called "IPv6: Moving to the Next Generation Internet."

Jeff Doyle has designed or assisted in the design of large-scale IP service provider networks throughout North America, Europe, Japan, Korea, Singapore, and the People's Republic of China. He is one of the founders of the Rocky Mountain IPv6 Task Force and is an IPv6 Forum Fellow.