An HPC Specialist Targets the Data Center
August 5th, 2009 By: Rich Miller
A major player in the high performance computing market is expanding its focus to the data center. Cluster Resources, whose Moab software is used to manage and automate the largest supercomputer environments, is introducing an expanded product suite for commercial data centers.
The move reflects the fact that virtualization and cloud computing are making many enterprise data centers look more and more like supercomputers. Cluster Resources is also adopting a new name and brand, Adaptive Computing, to reflect the broader focus of its products. In making the market transition, Adaptive Computing is depending its relationship with two key partners, IBM and HP.
High performance computing (HPC) employs supercomputers and computing clusters to manage advanced workloads. Historically rooted in scientific research in academics and government, use of HPC-style cluster computing has expanded into the energy and financial services industries.
From Supercomputer to the Data Center
Moab software manages workloads and virtualized resources for cluster, grid and utility computing environments. Moab currently runs in 12 of the world’s 20 largest computing installations, including Department of Energy laboratories at Los Alamos and Oak Ridge that house the world’s top two supercomputers, Roadrunner and Jaguar.
“Our software has been managing multi-billions of dollars of hardware,” said Peter ffoulkes, vice president of marketing for Adaptive Computing. “We’ve done well in establishing ourselves as the system of choice in high performance computing. Recent trends have made the traditional data center look more like an HPC environment.”
Adaptive Computing president Michael Jackson says this trend has been developing for several years. “In 2007, data center and cloud became about 50 percent of our business without us fully recognizing what was happening in the marketplace,” said Jackson.
But the HPC and data center markets differ in one key aspect - they often are managed by different teams. “HPC is typically on its own in a separate organization,” said Jackson, who said the company’s new offering will make Moab more accessible to C-level executives making decisions about data center infrastructure.
The new Moab Adaptive Computing Suite platform for data centers and private clouds is designed to allow a wider range of organizations to use Moab’s automation capabilities. New features allow Moab to work as a virtual data center-level operating system whose underlying infrastructure behaves like a shared pool of resources. The software can reconfigure systems and optimize the allocation of application services according to organizational policies and required service levels.
“The HPC community has long exploited the tremendous value of intelligent automation,” said Joe Zhou, senior analyst at Ideas International. “In this new age of enterprise computing, characterized by virtualization and cloud computing, a similar intelligence layer is required for optimizing resources in response to a dynamic business climate. Data center-wide virtualization and workload management can help to maximize the return on investments in infrastructure, applications, and personnel.”
Integrates with HP, IBM Tools
Moab integrates with leading data center management and automation platforms, such as IBM’s Tivoli and HP products based on Opsware automation. “Our technology does not displace these,” said ffoulkes. “It sites one layer higher and adds management capabilities.”
Jackson said that HP has used Moab in deploying a cloud computing infrastructure for the Defense Information Systems Agency of the U.S. Department of Defense, while IBM has combined Moab and Tivoli to deploy their private cloud infrastructure.
With the rebranding, Cluster Resources will continue as a business unit focused on the HPC market, while the Adaptive Computing unit will focus on the commercial data center space. Both will operate under the Adaptive Computing Enterprise parent company.