In addition to anything else we may be running out of, address space on the Internet — or at least the Internet Protocol we are mostly using now — is getting used up. As of June, 85 percent of the 4 billion individual addresses were being used.

Let me explain. Since the long ago development of nascent networking, the Internet has mostly used Internet Protocol version 4 — or IPv4. It uses a 32-bit address space that is capable of accommodating the 4 billion individual addresses. While that seems like a lot, the proliferation of computers and other networked devices such as digital video recorders, phones and set-top boxes, has been rapidly chewing up those addresses around the world.

Stopgaps to using up the space have included Network Address Translation (NAT) and private IP address spaces, which are effectively little bubbles of "hidden" networks behind a translating device to the Internet. But even these have only slowed the pace at which address space is being used up.

The solution adopted by the Internet Engineering Task Force (IETF), the Internet standards body, is IPv6. It is a revised protocol that supports 128-bit addresses providing up to 340 undecillion — that’s nearly 4 billion to the fourth power — addresses.

It is a substantial re-working of Internet protocol. While IPv6 incorporates a lot of great new features that should make networking more efficient and generally "better," it is not without costs. For it to work, everything needs to support it and, of course, that will take time.

Almost all newer operating systems — Windows XP and Vista, Mac OS X 10.3 and Linux, for example — come with built-in support but there are other things hanging off the Internet and migrating will take a while.

In fact, there has not been any great rush to migrate to IPv6. Over the last five years, I have done projects with federal government organizations with the specter of IPv6 looming in the background. An executive order mandating the government migrate to IPv6 by the end of June 2008 slipped because the expense of replacing so much infrastructure and re-working operations is so vast.

Now it is generally expected the transition to IPv6 will be like the original spread of IPv4, happening organically over time.

In the meantime, one can operate one type over the other through encapsulation so the transition can occur without too much pain.

The question at this point for business owners is: Why should I change since I can mostly see costs, not benefits, and my current IPv4, even with NAT, seems to be doing fine?

The answer is that more IP addresses, IPv6 has some nice features such as incorporated multicasting, true individual device-to-device communication, auto-configuration (kind of a mandated Dynamic Host Configuration Protocol allowing devices to be added to a network with little or no manual configuration), Internet Protocol Security (IPSec), Domain Name System security, flexible protocol extensions, and slowing down growth of the routing table, the master list of network destinations used by backbone Internet providers to communicate from one network to another.

I would recommend setting aside some time for your network folks to experiment and familiarize with IPv6 on a subnet. That way they can build experience with it and evaluate how it works. It allows your business to more judiciously plan how and when you might want to expand your use of it.

Some large Internet providers have already taken the plunge, including Comcast which has switched its core network to IPv6. Another source of a push may be from gamers into low-latency issues (the usually unnoticeable delay between input and output) and peer-to-peer file sharers because IPv6 works better for those applications.

 

Contact Lee LeClair, a founder and chief technology officer of Ephibian, through the company’s website www.ephibian.com or (520) 917-4747. Ephibian, headquartered at 3180 N. Swan Road, provides software development, data integration and Web design services. LeClair’s Tech Talk column appears the third week of each month in Inside Tucson Business.