Cray was recently awarded contracts for Cray Urika-GX system from customers specializing in manufacturing, and customer engagement. These new customers include the High Performance Computing Center of the University of Stuttgart (HLRS), and Phizzle - a customer engagement marketing and Internet of Things (IoT) company that enables the ingestion, analysis and reaction to Big Data in real time. The new contracts follow a previously announced collaboration with Deloitte Advisory Cyber Risk Services to offer threat analytics services powered by the Cray Urika-GX system.
"Cray has a long history of partnering with customers as they face complex and evolving analytics challenges", stated Ryan Waite, Cray's senior vice president of products. "The steady adoption of Urika-GX from a broader set of customers validates the benefits from integrating supercomputing hardware with powerful, open source analytics software like Apache Spark, to face those challenges. Our December update will provide additional support for enterprise customers with broader storage connectivity and broader data visualization support."
At HLRS, a project is underway to combine high performance computing and high performance data analytics to expand the Center's capabilities in automotive manufacturing. HLRS, which works in close collaboration with its industrial partners in the aerospace and automotive industries, will acquire two Urika-GX systems to help drive more predictive maintenance by implementing IoT style analytics with complex, large-scale modeling workloads.
"In the automotive industry, high performance computing and data analytics play an important role in product development, and with the progress of digitization, even larger and more complex data sets are generated that cannot be analyzed using conventional methods", stated Prof. Dr. Michael Resch, director of HLRS. "In cooperation with Cray and industrial users in the region, we are testing the possible applications of hardware in industrial environments, and the Urika-GX system will play an important role for us as we work to find a practical solution."
The collaboration between Cray and Phizzle will allow software developers on its phz.io platform to support customers at greater scale and speeds by achieving a 16x performance increase over their existing infrastructure. Developers will be able to leverage Phizzle's phz.io platform on its Cray Urika-GX system to gain supercomputing speeds that have previously been too complex or expensive to access.
"Powered by the Cray Urika-GX, phz.io will change the way everyday developers interact with their consumers", stated Stephen Goldberg, CTO of Phizzle. "By putting supercomputing and analytics in the hands of mobile, enterprise, and application developers who traditionally would not have access to Cray, phz.io allows developers to interact and react at speeds never seen before."
Cray is also introducing new software features to further enable enterprise customers. With the upcoming software release in December, Cray Urika-GX customers will be able to take advantage of a broader range of enterprise storage, including General Parallel File System (GPFS) and Network File System (NFS). Additionally, the Cray Urika-GX system will enable customers to leverage Tableau visualization tools and the latest version of Spark 2.0.
The Cray Urika-GX system features Intel® Xeon® E5 v4 processors, up to 22 terabytes of DRAM memory, up to 176 terabytes of local SSD storage capacity, and the high speed Aries network interconnect, which together provide the unmatched in-memory compute and network performance necessary to solve the most demanding big data problems. An exclusive feature of the Cray Urika-GX system is the Cray Graph Engine for fast, complex pattern matching and iterative discovery. With the Cray Graph Engine, customers can tackle multi-terabyte datasets comprised of billions of objects to uncover hidden relationships in even the nosiest of data. The Cray Graph Engine can run in conjunction with open analytics tools such as Hadoop and Spark, enabling customers to build complete end-to-end analytics workflows and avoid unnecessary data movement.