There was a great television commercial decades ago for Barney's, a fancy New York men's store. A bunch of kids, each of whom obviously represents a famous man as a child (Bogart, Stengel, Armstrong, LaGuardia), are sitting on a stoop. The group takes turns asking what the others are going to be when they grow up. The appropriate answers come back. Finally, one turns and says, "What about you going to do, Barney?" He looks at the camera. "I don't know," he says, "You're all going to need clothes."
Companies that supply bandwidth and Internet protocol version 6 (IPv6) technology are young Barneys. Converged applications, Web 2.0, unified communications and other tools are exciting. They will go nowhere, however, without an expansive underlying network. Vendors that supply the boring underpinnings of these networks will retire to very big homes in very nice places.
Perhaps the most fundamental element is Internet addresses, and IPv6 is necessary to replenish the supply. The BBC, in a story about the implementation of IPv6 on root servers, notes that the reservoir of IPv4 addresses is down to 14 percent, and is thought likely to reach exhaustion by 2011 if no action is taken. IPv6 massively increases available addresses by lengthening the string of numbers used to create them.
There are several reasons that IPv6 has lagged, however. The transition is pervasive and time-consuming, the current workarounds, such as Network Address Translation (NAT), create a false sense of security, and IPv6 itself (as opposed to the applications it enables), generates no new revenue. Experts also know that when it is deployed, IPv6 invariably will make their lives difficult. For example, this ComputerWorld post says that One Laptop Per Child PCs are outfitted both for IPv4 and IPv6. This can cause a conflict. It is a problem that, once recognized, can easily be fixed. However, it is certain to challenge help desks and lead to the kind of annoyances that IT staffs clearly will put off as long as possible.
The resistance to IPv6 seems deep. The story alludes to a survey from BT INS that indicates 58 percent of IT professionals aren't switching and that about 25 percent are not concerned about IPv4 address depletion. This long Network World piece concludes that federal officials are doing the least they can to comply with rules from the Office of Management and Budget (OMB). The bottom line is that the rules mandate only that federal networks support IPv6; they don't require that they actually be implemented. The story says that 10 percent of federal agencies are buying IPv6 services. The remainder are fulfilling the mandate upgrading their routers to be IPv6-capable, but aren't running the traffic.
There are some signs that people slowing are starting to deploy IPv6. The BBC story says that last week some Internet root servers began operating with a small number of IPv6 addresses. This means that for the first time two stations connected by the public Internet can find each other without using any IPv4 technology at all.
There also was a service introduction last week. Extreme Networks said that it is helping Bucks County Community College in Pennsylvania offer Internet2 applications and video collaboration. The underlying network will run at 10 Gigabits per second (Gbps) in the core and 1 Gbps at the edge and offer IPv4 and IPv6. The release from Extreme doesn't focus on IPv6, but it is clearly is a salient element of the a network that will support an exploding number of end points going forward.
The big question going forward is how quickly IPv6 will be deployed. This CircleID commentary asks whether vendors pushing the addressing scheme will be taken seriously if they are not using it themselves. The writer begins with a look back at Cisco's VoIP efforts. The vendor initially didn't use the technology internally, but soon recognized that it would not succeed in the marketplace until it did.
The question is whether the analogy holds for IPv6. The answer is that a vendor not using IPv6 sends a terrible message if there is a need for it. In many cases, however, the vendor's use case is far different than potential customers', and not using it can where it doesn't make sense can be justified.
The case for IPv6 is clear. However, there are good reasons to delay and do the minimum necessary. It will be interesting to see whether IPv6 will gain new momentum as the IPv4 address reserve continues to shrink.