How machine learning and predictive analytics drive business value

This chassis leasing company is deploying technology to proactively monitor its IT environment.
Written by Bob Violino, Contributor

Predictive analytics and machine learning are two technologies we're hearing a lot about these days. But do they actually present real opportunities for business value?

For Flexi-Van Leasing, one of the largest chassis providers in North America with 12 locations and a fleet of more than 130,000 units, the answer at this point is an emphatic "yes."

The company's chassis units are leased by trucking companies to transport cargo containers to and from ships and trains as part of the intermodal transport network. It's a deadline-driven environment in which any infrastructure downtime can backlog operations and give trucking customers a reason to bolt for competitors.

Flexi-Van is addressing this challenge by using the automation and predictive capabilities of the InfoSight Predictive Analytics framework from Nimble Storage. The company began using the technology about three years ago when it needed to expand data storage space for VMware VDI hosting, says James Mercer, CIO.

The company's IT infrastructure supports millions of transactions each day, and InfoSight helps it manage this environment more effectively, Mercer said, especially in terms of resource utilization and the monitoring of virtual machines (VMs).

The InfoSight platform collects and analyzes hundreds of billions of sensor data points in more than 9,000 customer environments. It then uses predictive analytics to solve any IT-related issues that might crop up.

Among the key benefits Flexi-Van has seen from using InfoSight are the identification of issues before users are affected. More than 200 Flexi-Van employees depend on applications within the company's infrastructure to run daily operations for leasing, billing processes, email, and other applications. InfoSight continually monitors infrastructure resources and identifies issues before users are affected, so that productivity remains intact.

Another benefit is the diagnosis of leading indicators that can impact performance. Flexi-Van's IT team uses InfoSight to identify and diagnose the top 10 "resource hogs" within its infrastructure. They can then reconfigure affected VMs to expand capacity and enhance performance as needed.

InfoSight also provides the company with a history of analytics, so it can understand how platforms have been performing and can use the data to accurately forecast future capacity, performance, and bandwidth requirements.

Unlike other storage platforms that can only capture and analyze a limited amount of historical data, performance data is sent to Nimble's cloud constantly, Mercer said. This data is then automatically used to examine and project trends, which are used to notify Flex-Van of impending resource requirements.

"From a senior IT executive perspective, the ability to do capacity planning well in advance makes the process of proposing expenditures much more timely, and better controlled," Mercer said. That, in turn, enhances the likelihood of getting major capital expenditure approvals.

From a technical resources perspective, the uptime and upgrade transparency makes for a trouble-free solution for data storage. "Being able to 'drive looking through the windshield instead of the rear-view mirror' is incredibly valuable to us," Mercer said.

Flexi-Van sees other potential uses of machine learning.

"I suspect there are hidden trends in the data we receive, trends that could help us predict changes in market volume of imports and exports in advance, by region and geographical distribution," Mercer said. "Machine learning could provide a way to automatically find those trends, extrapolate from them, and notify us of potential upcoming changes in various locations."

Read more:

Editorial standards