2018 Will Bring More Predictive Maintenance to IT Assets
Park Place Hardware Maintenance
The nirvana being sought is the autonomous data center in which AI-driven infrastructure management (DCIM) software “will monitor and control IT and facilities infrastructure, as well as applications, seamlessly and holistically.”
An average of 1.6 hours of downtime each week, at a cost of at least $138,000 per hour. These statistics aren’t describing the email server at an understandably cash-strapped charity. This is what over half of Fortune 500 companies experience, according to Data Center Dynamics.
Often the impacts of downtime on businesses can be much worse. Other estimates put average downtime losses at nearly $900,000 per week for a company with about 10,000 employees. A 2016 data center outage at Delta airlines cost $150 million in dollars and cents and had untold impacts on its reputation and customer relationships.
The Rise of Predictive Maintenance
Data centers are looking hard at predictive maintenance to help avoid these costs associated with network downtime. They’re taking a page from the “industry 4.0” or industrial IoT (IIoT) playbook with techniques shown to work in factories, along oil pipelines, and in other industrial settings, Why not in tech facilities, too?
The trend is a departure from reactive maintenance, which responds only when problems arise. It’s also a step above scheduled maintenance, which tries to prevent downtime with regular interventions timed based on historical lifespan data.
Predictive maintenance uses real-time data to anticipate issues before they happen. In fact, a computerized maintenance management system (CMMS) can identify potential issues and launch trouble tickets with no user intervention.
Park Place recently debuted the ParkView infrastructure managed services, a remote triage service platform that enables predictive detection and identification of hardware faults that occur within a data center.
ParkView, powered by BMC technology, revolutionizes visibility into data center infrastructure and operations by identifying and reporting hardware faults, as well as potential faults, enabling faster response and problem resolution. ParkView predicts data center issues, then triages the fault and identifies the proper fix, allowing quick repairs to be made through Park Place’s seamless integration with hardware maintenance service plans.
Better Facilities Maintenance
There are clear applications to data center facilities. Take an example as simple as an air conditioning filter. Standard maintenance would change out filters on a scheduled basis to help prevent an A/C failure on a hot day bringing down dozens of pieces of hardware.
But filters clog more quickly in open-air data centers, especially those close to urban centers, than they do in sealed facilities. Other factors can affect timing as well. Predictive maintenance one-ups scheduled maintenance by acknowledging the fact and using certain readings—perhaps air flow rate—to determine when the filter is clogging and should be changed.
The nirvana being sought is the autonomous data center in which AI-driven data center infrastructure management (DCIM) software “will monitor and control IT and facilities infrastructure, as well as applications, seamlessly and holistically.” Cooling, power, networking, compute workloads, storage, etc., would be managed dynamically.
Getting Predictive in 2018
It’s fun to imagine a future in which dexterous robots replace drives at the earliest sign of failure, based on analysis of information provided by zillions of sensors strategically placed around the data center. But this level of automation is still a ways off.
Predictive maintenance is, however, having a big impact and will be copied by more cloud operators and on-premises data centers in the coming months.
FIIX CMMS says a move from corrective maintenance to preventive maintenance delivers savings of 12% to 18%, on average, and the next step to predictive maintenance adds about 12% more savings. How can any data center operator pass that up? Successes like Google’s—a 40% reduction in energy used for cooling and a 15% reduction in overall energy overhead—will serve as proof of concept and drive interest in and adoption of CMMS.
There are additional ways to integrate a predictive mindset into IT maintenance, beyond facilities management. High-level third party maintenance providers offer remote monitoring for storage arrays, servers, and networking equipment. They keep an eye on the systems and can intervene proactively if real-time information shows a problem may be forthcoming.
It’s a great way to tip the balance toward a proactive stance and cut downtime as well, at least until the robots can take over.
Chris Adams is President and COO of Park Place Technologies. Contact him at firstname.lastname@example.org.