The Magic of IPv4

Early in my network career I had to explain to my boss why we needed head-end routers in the data center. Coming from an application background, he didn’t understand data connectivity and thought the packets somehow went directly into the mainframe through the “ether” net I kept talking about. Somehow I survived that career-challenging meeting and adequately explained basic network connectivity in a way that allowed me to keep my job and purchase the needed routers.

Looking back, there were times when it seemed that there was some magic to data connectivity. Our ability to access files and resources on a computer in another part of the world over telephone lines caused many people to call us network wizards. And with a lexicon of magical terms such as DECNet, TCP/IP, SNA, NetBEUI, ASCII and IPX, we conjured up connectivity solutions that verged on miraculous (it was often a miracle they worked at all!). However, after 22 years of “conjuring,” I can safely say there is nothing magical about networks — with the possible exception of IPv4.

By all reports, IPv4 should be dead. Its limited address space, antiquated security and lack of native authentication mechanisms have generated multitudes of obituaries by industry pundits who proclaimed the new era of IPv6. According to the Merlins of Internet connectivity, IPv6 was required for the widespread adoption of VoIP, B2B extranets, and secure Internet connectivity.

To continue reading, register here and become an Insider. You'll get free access to premium content from CIO, Computerworld, CSO, InfoWorld, and Network World. See more Insider content or sign in.

Early in my network career I had to explain to my boss why we needed head-end routers in the data center. Coming from an application background, he didn’t understand data connectivity and thought the packets somehow went directly into the mainframe through the “ether” net I kept talking about. Somehow I survived that career-challenging meeting and adequately explained basic network connectivity in a way that allowed me to keep my job and purchase the needed routers.

Looking back, there were times when it seemed that there was some magic to data connectivity. Our ability to access files and resources on a computer in another part of the world over telephone lines caused many people to call us network wizards. And with a lexicon of magical terms such as DECNet, TCP/IP, SNA, NetBEUI, ASCII and IPX, we conjured up connectivity solutions that verged on miraculous (it was often a miracle they worked at all!). However, after 22 years of “conjuring,” I can safely say there is nothing magical about networks — with the possible exception of IPv4.

By all reports, IPv4 should be dead. Its limited address space, antiquated security and lack of native authentication mechanisms have generated multitudes of obituaries by industry pundits who proclaimed the new era of IPv6. According to the Merlins of Internet connectivity, IPv6 was required for the widespread adoption of VoIP, B2B extranets, and secure Internet connectivity.

After years of defending IPv4, I finally caved into the overwhelming evidence being provided by industry analysts, Internet experts and my own engineering staff. With arms wide open I stood ready to embrace the future and begin migrating my networks to IPv6.

And now two years later, my arms have grown tired of waiting for something to embrace; my engineering staff has moved on to other jobs; and my network is supporting VoIP, IPSec, SSL, extranets, intranets secure remote access and everything else with antiquated IPv4.

My IPv4 network-addressable devices are increasing daily, stricter security requirements are being mandated monthly, business-critical voice and data applications are being deployed weekly. I’m communicating with external devices that use the same 10.0.0.0 address space that my networks use, and my NAT appliances are not the traffic bottlenecks they were forecasted to be. How can that be?

Some will say this is due to the technology advances such as ASICs, faster backplanes, dedicated routing engines and more efficient software that enables IPv4 to overcome its inherent limitations and provide business-critical network connectivity. They will also say that these are interim measures and we still need to be preparing for an eventual migration to IPv6.

That may be, but maybe my old boss was right. Maybe ARPA knew more than we realized. Maybe the Department of Defense did indeed have alien assistance. Maybe there’s more to IP subnets than bit-level masking. Maybe there’s more magic in IPv4 than we realize … (cue X-Files theme song).

Read more about infrastructure management in Network World's Infrastructure Management section.