The secret sauce of the Internet is “IP” (short, of course, for Internet Protocol). IP is the fabric upon which the whole Internet ecosystem is based. Fortunately, as complex as the Internet Protocol is, it’s largely transparent to end users. But as network engineers are aware, the predominate version of the Internet Protocol in use today is version 4, commonly written as IPv4. However, IPv4 can only support 4.2 billion unique addresses, a painfully small number given the explosive growth of the Internet and IP connected devices. IPv6, the version to succeed IPv4, allows for a nearly infinite number of addresses.

It appears that IANA, the International Internet authority that doles out addresses to the regional registrars, will more than likely run out of IPv4 address allocations to the registries somewhere in the late '09 or early '10 timeline. However this does not imply that there won't be any space left to distribute. Most operators that we have spoken with place a realistic 'red zone' date in the 2011 time frame. Of course IPv4 is not going to just miraculously disappear or be mass converted to all v6 addresses within the next decade. We will be dealing with the support of v6-only devices, v4-only devices and dual-stack devices for a very long time. I asked one of Cox’s ace engineers, Jason Weil, to summarize what this will mean for cable operators, and here's what he told me:

- IPv6 will need to be extensively tested in the MSO labs, both in a native deployment, as well as in a dual-stack environment.

- network tool support for IPv6 will be the first major challenge for operators to overcome. This will include fundamentals, such as DNS & DHCP software and hardware platform development and testing. Support for v6 IP provisioning will be a large part of this work due to significant lab validation efforts.

- a protocol translation (6-to-4 and vice versa) service platform will have to be designed or outsourced to allow for intercommunication of IPv6-only to IPv4-only hosts for the decade or longer transition period.

- standards still need to be developed and others fine-tuned in both the IETF and CableLabs realms to deal with problematic areas of operation and integration including security standards, multi-homing, and protocol selection.

- the residential OS situation will also be a major obstacle to overcome. Vista is the first true Windows OS that supports IPv6, and applications that support IPv6 running on Vista are even farther out on the horizon.

- we will need to prepare for the inevitable customer education issues that will arise from this change. Customer support and communication strategies will need to be developed.

The good news is that there is awareness of the issues and work has begun to address them. Some Asian operators, responding to government mandates, have a jump start on this transition, so we certainly can gather lessons learned from our international peers. But this is no longer a topic where we simply grimace and push off the pain to a future date. The planning and work needs to start now, so we can ensure a smooth and transparent transition.

Jay Rolls is Cox's VP of Technology