Western Digital Brings Analytics To R&D Processes As It Reinvents The Hard Disk Drive

Most organizations use analytics software to better understand their customers or refine their sales strategy. Western Digital is using it build the next generation of hard drives that will be used in businesses all over the world.

Based in San Jose, Calif., Western Digital has already managed to achieve a number of world firsts, including the first disk drive and, last year, the world’s first 14TB hard disk drive (HDD). According to Gautam Khera, however, the next wave of innovation is coming through using analytics to develop and refine how HDDs are manufactured using elements that range from helium to microwave energy.

“HDDs are the coolest high-tech gadget next to the human body,” the firm’s senior global director of technology R&D told the Analytics Experience conference produced by SAS in San Diego this week. “This device has solid state components like the chip, mechanical moving parts like the motor, and chemical components like the air filter.”

As sophisticated as they are today, Khera said Western Digital is being driven to work harder than ever to meet the increased data storage expectations of business and consumer users. This is a highly complex task, however.

Today’s HDDs, for instance, are made up of more than 250 components involving work across 12 industries, 150 suppliers and 300 locations. This includes processes around casting, foraging, machining and high-precision assembly.

One of the biggest areas of opportunity is to enhance the magnetic recording head, Khera explained. Much of the industry has been focused on the potential of heat-assisted magnetic recording (HAMR), but it requires an expensive laser to lower the energy barriers involved, which can lead to lower reliability. Instead, Western Digital has been developing HDDS based on microwave-assisted magnetic recording (MAMR), which uses an electrical magnetic field to allow data to be written by the storage media.

Working on MAMR-based HDDs starts with a ceremaric wafer that goes through 900 different process steps, building components one layer at a time.

“We are generating lots of data in each process step,” Khera said. It still takes a long time to build it, though: up to nine months on average to create a device and get the necessary feedback from various stakeholders.

Analytics tools such as SAS’s software is allowing Western Digital to predict what might happen in the various steps during the manufacturing process, from the creation of a wafer test chip to the test itself and beyond.

“Traditionally it’s been a very long cycle of learning. We need to improve that,” he said, adding that in some cases analytics has already saved weeks worth of time.

Even if its application of analytics is somewhat unusual, Khera said Western Digital has had to be thoughtful about the strategy behind it, such as simplifying processes.

“In R&D, the end to end data sets are the critical data asset. We have to make sure they’re not just sitting in a silo with one engineer,” he said. “We need to assemble all the data from the process steps into a usable, easy to access place.”

It’s also important to foster a culture where disparate parts of the business can work cohesively, he added. Analytics projects can span multiple business units and centres of excellence as well as the R&D group, he said.

“We want to create advanced analytics trust across the organization,” he said. “We need to do all this together — and to speak the same language with common analytics tools.”

SAS’s Analytics Experience wrapped up on Wednesday.