typewriter keys

The IT department shouldn't have to worry about the state of its network’s infrastructure, a storage systems company executive told me just yesterday.

If technology is doing its job properly, then IT should be able to focus on delivering customer experiences, and let the dirty business of network infrastructure handle itself.

The state of Internet Protocol network infrastructures worldwide may not yet be generally deteriorating. But it has reached the point where, in most countries on average, vendor support for more than half of the world’s network devices, switches and routers is no longer available.

By Cisco’s standards, that means these devices are typically six years old or older. If the IT department isn’t looking after these devices, nobody is.

This finding is the key takeaway from the 2015 Network Barometer Report, released this week by data center connectivity provider Dimension Data. The survey assesses the network infrastructure hardware in use by its own clients worldwide.

The Foundation Still Matters

Network infrastructure is the support system underlying all the applications that collectively comprise the “customer experience.” In a great many customer surveys asking what vendors can do to improve the experience, their responses center around speed.

An e-book on customer experience management published just last month by research services provider Qualtrics [PDF, registration required] suggests that the pace of innovation has built up greater expectations among consumers for faster response times. “Speed in data collection, analysis and action,” the e-book notes, “can be the singular difference between a company rising to prominence or becoming irrelevant.”

Now, you could have the IT department ignore infrastructure issues the same way people put off colonoscopies, and devote its attention to “what it does best:” the surface issues, such as where which box should go where in the layout. Or, as Dimension Data suggests, you can keep your systems up-to-date and running smoothly.

Tolerated Obsolescence

Just four years ago, North and South America led the world in maintaining newer network infrastructure as of 2011, with only 22 percent of network equipment having aged five years or more. But this year, the Western hemisphere is stuck with the oldest equipment of all. Some 60 percent of network equipment has aged to the point where Cisco and Dimension Data considers it “obsolete.”

Some 16 points of that jump came in just the last year, although Dimension Data admits the principal cause of that jump came from “a single assessment conducted for a large organization in the government sector.”

Keep in mind, this does not mean “non-functional” or “worn out.” Cisco is believed responsible for a majority of the world’s networks, and arguably, its devices are built to last. What it does mean is that the platforms supporting at least half of the world’s connectivity are, at best, being kept on life support until they can be replaced.

But that replacement time actually does come soon, according to Dimension’s researchers. The pattern appears to work like this: Companies bleed the last ounce of supportable product life out of their hardware, with the intention of replacing it immediately thereafter. Sometimes, companies never find the time to act on their intentions.

Why This Isn't a Hardware Story

As you might imagine, one of the services Dimension provides is device lifecycle management, where it reduces the percentage of obsolete devices in network infrastructures in active service to around 5 percent. Arguably, that’s one way to alleviate the burden of thinking about the systems that sustain their business, from the IT department and the CIO.

But in the small print of an appendix that might otherwise have been mistaken to have read, “This page intentionally left blank” (along with a list of reasons why), Dimension had the presence of mind to mention another possible solution, which the company is investing in.

“Software-defined networking makes networks more intelligent, programmable, and automated,” the footnote reads. “It’s brought about by changes at the networking device level.”

One of those changes, had Dimension given itself the time and space to explore it, is enabling IT to adopt an up-to-the-minute deployment model for network infrastructure, using the same continuous deployment model being used for customer-facing applications.

Last year, when the Heartbleed bug affected the security of Web sites everywhere, the cause was traced to a long-overlooked defect in the way the OpenSSL open source encryption library was written. That defect was corrected, to some degree, by patching the firmware of network devices that utilize OpenSSL. But not every piece of hardware in the network can be so easily patched.

If the platform upon which networks ran were software-based, theoretically not only could the OpenSSL bug have been corrected as soon as it was noticed: but since there would be a continuous deployment model in place, developers and engineers would have had more incentive to hunt down, isolate and correct such bugs more often than once every five years.

Put another way, if the technology lifecycle for fundamental platforms were the same cycle as for user applications, the entire systems management process could focus its efforts on improving customer experience … rather than just the web team.

Simpler Media Group, 2015