Brian Wilson (pictured) is senior vice president of customer success at Zenoss, a solution provider based in Austin, Texas. He can be reached at bwilson@zenoss.com.

In my discussions with CIOs and other IT leaders, I am frequently asked one question: Does networking monitoring really matter now, with cloud technology maturing so rapidly?

The simple answer is yes. In fact, network monitoring matters more than ever.

Here’s why. As fundamentally disruptive as the cloud is to IT, one fact remains: Until at least the end of this decade, most enterprise applications will continue to be run on traditional data center infrastructures.

Let’s look at the numbers. In a report last summer, analyst firm IDC determined that the cloud represents one-third of all new IT infrastructure spending. That’s expected to increase through 2019, IDC says.

While that statistic is eye-opening, the converse is equally interesting:

•    67 percent of IT spending is on non-cloud based infrastructure, and

•    $67 billion was spent in just the last year on new non-cloud infrastructure.

That’s a tremendous amount of infrastructure that’s subject to the increasing availability and performance demanded by service-level agreements. And the money spent on the cloud in many cases works with, not independently of, the physical network.

As badly as the C-suite wants a move “all in” to cloud, the business case to make that move varies widely. New applications built with the cloud in mind are a no-brainer, but what about legacy systems running the back end of the business? Is it worth the investment to migrate to the cloud

Likely not.

Cloud migration does not mean simply moving an application from here to there; it means re-architecting that solution to take advantage of ephemeral services, streaming data and distributed components. That means time and money. Most businesses simply cannot afford to toss out the investments of legacy applications or to invest even more to move everything to the cloud.

At the same time, the cloud has placed increased expectations of availability and performance on legacy applications. Availability of 99.9 percent and response time of 80 milliseconds were once acceptable. Now, new generations of solutions built for the cloud are driving businesses to look for ways to cut response times in half or to add fractions of a decimal to availability.