Managing Change Key to Blending HPC, Big Data Operations...
When it comes to finding ways to maintain an edge in the competitive storage market, HGST, a Western Digital company that initially spun out of IBM and Hitachi, is counting on the cloud. But for HGST CIO, Steve Philpott, the initial challenge wasn't the technology — it was implementing the roadmap to increased flexibility and cost savings.Philpott led the “zero to cloud" charge at HGST in just over six months, helping the company move critical components of its daily operations, including its sales and internal business practices, to the cloud before. It was a prelude to a much broader initiative — one that would enhance the agility of their teams on the research and development front for a range of big data and high performance computing applications.From his experiences leading another company into the cloud frontier at Amylin Pharmaceuticals, which was one of the first companies in its field to move mission-critical, regulation-sensitive workloads into a virtualized environment, Philpott knew the speed, flexibility, and potential for optimization that would be possible for the diverse teams at HGST. However, his toughest job was convincing teams that had always been reliant on on-premises clusters that performance remain strong without adding more time and cost to their operations."This is probably the most exciting time to be in IT because we now truly have the tools and technology to enable business to move at a pace it could never move before," Philpott explained. However, he notes that in the end, even with a vast new opportunity on the tools front, the technology isn't the hardest part — it's the change management process that is the real challenge.For example, he explained how HGST's research division in Japan initially had a hard time seeing outside of the on-site HPC cluster. But with the relative ease of spinning up test environments to run compute-intensive molecular and fluid dynamics applications, they saw the performance and optimization capabilities of running inside Amazon Web Services (AWS). Their conclusion: Now they could focus their time on solving problems versus managing hardware.Within 16 weeks, the company's HPC division, including demanding high performance computing applications, were handled in the cloud, eliminating the need for an outright purchase of a new cluster while bolstering the team's ability to change at the pace of new ideas.Further, the much-needed computational resources were up and running four months earlier than if the team had waited for a new on-premises HPC cluster.The change management process was extended to include the daily demands of big data — particularly on the rigorous production side. With its range of devices, including new low-cost, high-capacity sealed helium-filled drives, HGST needed a feed of manufacturing data from several sites meshed with a wealth of other data from across business units. This led to the development of an entirely cloud-based big data platform as pictured below, which was able to seamlessly blend and orient product goals and targets with practice — and to help HGST remain agile enough to scale and optimize around challenges and opportunities. What you see below represents the end-to-end “DNA of a hard drive" as addressed at all hops by the AWS Cloud. Aside from the diversity of tools handled within the big data platform, the company optimized their AWS environment on the storage side, taking full advantage of the range of tiers available to manage data use and storage using an HSM approach."We have probably 30 plus terabytes in EBS, 300 terabytes or more in S3, and 3 to 5 petabytes in Glacier," said Philpott. "The key is that it's there with three years of data always on the ready, but having the tiers lets us optimize both performance and cost."While the specifics of the big data and HPC platforms themselves are fodder for articles richer in technical detail, the real story of HGST's success lies in implementing a solid change management strategy. It was hard to get users to think outside the box. But once implemented, the company is leaner, more nimble, and able to accelerate their pace of innovation.When it comes to finding ways to maintain an edge in the competitive storage market, HGST, a Western Digital company that initially spun out of IBM and Hitachi, is counting on the cloud. But for HGST CIO, Steve Philpott, the initial challenge wasn't the technology — it was implementing the roadmap to increased flexibility and cost savings.Philpott led the “zero to cloud" charge at HGST in just over six months, helping the company move critical components of its daily operations, including its sales and internal business practices, to the cloud before. It was a prelude to a much broader initiative — one that would enhance the agility of their teams on the research and development front for a range of big data and high performance computing applications.From his experiences leading another company into the cloud frontier at Amylin Pharmaceuticals, which was one of the first companies in its field to move mission-critical, regulation-sensitive workloads into a virtualized environment, Philpott knew the speed, flexibility, and potential for optimization that would be possible for the diverse teams at HGST. However, his toughest job was convincing teams that had always been reliant on on-premises clusters that performance remain strong without adding more time and cost to their operations."This is probably the most exciting time to be in IT because we now truly have the tools and technology to enable business to move at a pace it could never move before," Philpott explained. However, he notes that in the end, even with a vast new opportunity on the tools front, the technology isn't the hardest part — it's the change management process that is the real challenge.For example, he explained how HGST's research division in Japan initially had a hard time seeing outside of the on-site HPC cluster. But with the relative ease of spinning up test environments to run compute-intensive molecular and fluid dynamics applications, they saw the performance and optimization capabilities of running inside Amazon Web Services (AWS). Their conclusion: Now they could focus their time on solving problems versus managing hardware.Within 16 weeks, the company's HPC division, including demanding high performance computing applications, were handled in the cloud, eliminating the need for an outright purchase of a new cluster while bolstering the team's ability to change at the pace of new ideas.Further, the much-needed computational resources were up and running four months earlier than if the team had waited for a new on-premises HPC cluster.The change management process was extended to include the daily demands of big data — particularly on the rigorous production side. With its range of devices, including new low-cost, high-capacity sealed helium-filled drives, HGST needed a feed of manufacturing data from several sites meshed with a wealth of other data from across business units. This led to the development of an entirely cloud-based big data platform as pictured below, which was able to seamlessly blend and orient product goals and targets with practice — and to help HGST remain agile enough to scale and optimize around challenges and opportunities. What you see below represents the end-to-end “DNA of a hard drive" as addressed at all hops by the AWS Cloud. Aside from the diversity of tools handled within the big data platform, the company optimized their AWS environment on the storage side, taking full advantage of the range of tiers available to manage data use and storage using an HSM approach."We have probably 30 plus terabytes in EBS, 300 terabytes or more in S3, and 3 to 5 petabytes in Glacier," said Philpott. "The key is that it's there with three years of data always on the ready, but having the tiers lets us optimize both performance and cost."While the specifics of the big data and HPC platforms themselves are fodder for articles richer in technical detail, the real story of HGST's success lies in implementing a solid change management strategy. It was hard to get users to think outside the box. But once implemented, the company is leaner, more nimble, and able to accelerate their pace of innovation.…
Share
TRENDING