Speeding up workloads: OpenPOWER takes aim at analytics and cognitive computing


All things in the enterprise come down to the data center. Data must be stored, presented, moved and processed. Most importantly, data must live somewhere, and it’s to everyone’s benefit if that somewhere can serve up the data as quickly as possible.

The OpenPOWER Foundation is a group built around using open standards to customize data center technology for the unique needs of business. By customizing their data center systems, companies can get the best speed possible for the workloads they run, according to Steve Roberts, offering manager for big data on Power Systems at IBM Corp.

“Ultimately, it’s about improving time to insights,” Roberts told Dave Vellante (@dvellante), co-host of theCUBE, SiliconANGLE’s mobile live-streaming studio, during the DataWorks Summit in Munich, Germany. (*Disclosure below).

Designed for machine learning

Roberts described the OpenPOWER Foundation as a group dedicated to innovating through IBM’s POWER CPU processor in all aspects. The objective was to help drive OpenPOWER into the market. They sought to marry the benefits of an open hardware ecosystem with the philosophy of the open software stack.

It’s no coincidence that OpenPOWER is built to run cognitive workloads. POWER systems has a server built to run big data and machine learning applications. Likewise, the technology will be part of the cognitive systems portion of IBM, Roberts explained.

The advantage to this system comes in the form of highly parallelized workloads, which cognitive computing is designed to handle very well. The system also runs accelerators that can improve processing for specific tasks, he added.

“Not everyone is going to start with souped-up accelerated workloads, but they’re going to get there,” Roberts said.

It’s hard to predict where workloads will go, so it’s good to start with a server that has room for growth, he stated, adding that the focus is being able to run analytics where they make sense for the company.

Watch the complete video interview below, and be sure to check out more of SiliconANGLE’s and theCUBE’s coverage of DataWorks Summit 2017 EU(*Disclosure: TheCUBE is a paid media partner at DataWorks. Neither conference sponsor Hortonworks nor IBM, sponsor of this segment, has editorial oversight of content on theCUBE or SiliconANGLE.)

Photo: SiliconANGLE