Stillborn IPv6 offers no compelling reasons to switch

Recently, I had a chance to dine with some of the top network implementers in the world: the people who put together NetWorld+ Interop's network twice a year. These folks have been through every advanced technology you can imagine. For them, Internet Protocol Version 6 (IPv6) is old hat - Interop showcased an IPv6 network years ago. "So when will IPv6 hit the mainstream?" I asked. Their answer: "Never."

IPv6, the next version of the venerable IPv4 protocol used in the Internet, was designed to solve the problem of address exhaustion. There's more to IPv6 than that, of course; for example, it includes IP Security virtual private network security as a built-in, rather than add-on, feature. But IPv6's fundamental design is built around the shortage of 32-bit integers assigned as IP addresses.

While IPv6 was going through its highly contentious and political development, the Internet engineers devised alternate techniques that help to reduce the need for globally unique IP addresses, including classless inter-domain routing (CIDR) and network/port address translation technologies. These alternate technologies were wildly successful. Thus, there is no longer a compelling need to run to IPv6 - and because of the high cost of switching to IPv6, there are great pressures not to change.

Globally unique and routable IP addresses have become a scarcer and much more expensive resource, and the Internet routing tables have exploded. At 62,000 routes, it takes a substantial amount of CPU power and memory to handle those tables. Hint: buy Cisco, Nortel Networks and 3Com stock - they're going to make a lot of money selling upgraded routers powerful enough to handle the Internet. In the meantime, organizations are hoarding IP addresses. Building a fault-tolerant Internet connection is expensive, complex and difficult.

This could be solved if we used IPv6. For example, if IP addresses started to become long and ugly, people wouldn't want to hold onto them and might start using the Domain Name System to solve some of their IP addressing issues.

It's easy to throw stones at Microsoft, which is not going to release IPv6 with Windows 2000 (and probably not with Windows 2001, either). But even if IPv6 did ship with Windows 2000 next month or next year, it wouldn't matter; organizations wouldn't use it. In the absence of a major crisis, why screw with a company LAN that's just barely under control today?

We're not going to go to IPv6 because it doesn't solve a lot of the other problems we are discovering in managing and maintaining the Internet - and because there is no longer a compelling reason to do so. But IPv4 does have a limited life. There will be some other protocol designed to solve the problems of the next century. And that protocol will eventually replace IPv4, first in the backbone infrastructure and then moving out to the desktop. In the meantime, chill out, learn your CIDR tables and prepare to buy some fatter routers.