No One Expects Squirrels: Why Data Center Maintenance Must Be Proactive

Forget about love. In the Digital Age, data centers make the world go ’round, and there’s a reason companies invest time and resources in ensuring availability: Downtime costs a fortune. When Amazon experienced 49 minutes of downtime in 2013, it cost the company $4 million in lost sales.

And it’s not just megabrands that are at risk. On average, organizations lose $138,000 per hour of downtime. With 59 percent of Fortune 500 companies experiencing at least 1.6 hours of downtime each week, it’s easy to see how data-center disruptions translate into real costs.

In addition to normal wear and tear, data centers face a daunting array of maintenance challenges. From natural disasters and cyberterrorism to human error and squirrels, the challenges data centers are up against are multiplying as the cost of downtime continues to rise.

Fix Tomorrow’s Problems Today

Data centers – whether your own or a customer’s site – need to be able to identify and remedy potential problems before they disrupt service. But across a range of industries, performance-minded firms are discovering that prevention isn’t the only game in town; prediction also plays an important role in eliminating the risk of disruptions. What’s the difference?

Preventive maintenance is a calendar-based approach that works to prevent disruptions by scheduling maintenance and upkeep activities on a weekly, monthly, quarterly or yearly basis, for example. Systems that feature preventive maintenance routines typically use historical data and manufacturer recommendations to develop maintenance schedules tied to expected machine life cycles.

Predictive maintenance, in contrast, uses real-time data to determine the actual status and condition of machinery and relies on monitoring systems to gauge conditions and trigger follow-up actions, like scheduling work orders to resolve concerns that arise outside of calendar-based, preventive maintenance cycles.

Although it can be useful to use preventive maintenance for less-critical system components, predictive maintenance is a more reliable approach for mission-critical systems — like the hardware that data-center facilities depend on to provide seamless availability and service to users. Still, despite the need for uninterrupted service, many data centers continue to rely on preventive-maintenance routines held together by a dangerous combination of electronic spreadsheets and paper charts.

So how do you sell this kind of service?

The primary point is the high costs associated with even a few minutes of downtime. As we discussed in our recent report on Internet of Things and MRR, predictive maintenance can eliminate the potential for unexpected problems between scheduled maintenance events. In addition, predictive maintenance can lower costs by reducing the frequency of scheduled maintenance tasks, freeing up technicians to focus their time and energy on more critical tasks.

Less tangible but just as important, predictive maintenance also protects the organization’s reputation. For many brands, even short periods of downtime can severely damage customer relationships and create a PR nightmare that could have been avoided. By incorporating a predictive approach into maintenance routines, partners that support data centers can take a big step forward in meeting the needs of customers in the evolving digital marketplace.

Paul Lachance is president and chief technology officer for Smartware Group, Inc., producer of the award-winning cloud-based Bigfoot CMMS. Lachance has been developing and perfecting CMMS for the maintenance professional for more than 20 years. Contact Paul directly at paul.lachance@bigfootcmms.com or visit bigfootcmms.com.