Few, if any, information technology subjects have garnered more attention over the past two years than Big Data. It's a savior. It's evil. It's the killer app of killer apps. And now, according to Gartner analyst Svetlana Sicular, "Big Data Is Falling into the Trough of Disillusionment."

This disillusionment is giving me a little bit of Big Data fatigue - at least in the sense that so much buzz is focused on what Big Data can do for marketing and business intelligence departments. Lost amid all the hype and hoopla is the story of how predictive analytics can help IT teams better understand and support the business applications that are the very lifeblood of their enterprise. After all, predicting the right service to offer your online customer won't help if the customer experiences sluggish application performance.

And Big Data is not going away. Big Data begets Very Big Data, and so on. Analysts are projecting that astronomical amounts of money will be spent on analyzing and using information. This includes projections that millions of data scientists are going to be hired to make sense of all this. You and I know that few if any of those data scientists will be monitoring application or infrastructure performance.

Application performance monitoring is a critical job that can substantially impact a company's brand and bottom line. The fact is that this is an incredibly tough assignment because of all the variables that can negatively impact performance. It is like telling an Ops pro, "You don't have to find a needle in a haystack; you have to find an unspecified number of needles in an unspecified number of haystacks." As databases and other apps get bigger and more distributed, this is almost impossible to do with human intelligence and traditional tools.

The good news is that predictive analytics for IT solutions are far simpler and increasingly more affordable than their Big Data counterparts. It is important for all of us to bang the drum louder about the business value that predictive analytics can provide for IT, not just for application performance, but for people performance too. The industry needs to discuss the risks of fielding powerful and complex applications without arming IT Ops and Support with similarly powerful tools to ensure their availability and performance.

Why Predictive Analytics for IT Is So Cool - and Worth Talking About
Most currently available analytical tools that can forecast human behavior give users exactly what they ask for - and that is part of the problem. The more variables you have in an equation, the more relationships exist. It's getting more difficult all the time to know what questions to ask.

The cutting edge of predictive analytics for IT uses advanced machine intelligence to scan data patterns to detect anomalies and other relationships that human analysts cannot see, at least not see quickly or without the help of data scientists. Using machine intelligence as part of IT performance management is like having your own personal R2D2, or an android like Star Trek's Lieutenant Commander Data. In fact, pairing machine intelligence with human intelligence is the future of problem solving and prediction - you can see it working today to identify root causes of performance failures and limit unplanned downtime. It is quite cool. There is hope for the disillusioned.

Additionally, the list of use cases for predictive analytics in IT is growing. We see organizations using it all the time to predict and prevent service disruptions. If something does go wrong, users can often find the source of the disruption much faster than they can with just human analysis. I have found that in some cases, machine intelligence can identify root causes of performance failures up to 90 percent faster than human intelligence alone can.

We can already see innovations on the near horizon that will make predictive analytics powered by advanced machine intelligence even better equipped to serve IT Ops' needs. Integrating performance management with security threat management is a good example. All of this is good news for IT Ops pros because their jobs are increasingly vital to an organization's success in this era of widely distributed infrastructures.

This is why companies have to emphasize the business value of predictive analytics for IT in conversations about application and infrastructure management. If you forget about IT Ops' needs when contemplating the deployment of data-heavy business applications, you put your network and your organization as a whole at risk.

Predictive analytics in IT is a breath of fresh air in the Big Data conversation. It gets me excited - far from being disillusioned.