GOVERNMENT

Penguin Computing introduced its new blade server, BladeRunner. When packaged with Scyld Beowulf Linux clustering software, BladeRunner becomes a "cluster in a box," the first offering of its kind. Scyld Beowulf is a complete, standards-based Linux distribution tailored for high performance computing (HPC) clusters. BladeRunner is also the ideal datacenter consolidation platform. This Linux optimized blade server offers twice as much compute density when compared to competing products the company said in a release yesterday. "The cluster-in-a-box package complements our already extensive cluster product line by making clustering a viable option for departmental HPC clustering applications and specialized research organizations who have to do more with smaller teams and limited support staff," said Enrico Pesatori, chairman and CEO, Penguin Computing. "With powerful products and industry knowledge, Penguin Computing helps customers reduce the time to market and increase staff productivity. This ultimately levels the competitive landscape by allowing customers to do more for less, without affecting performance or quality." Faced with ever-increasing performance requirements and budgetary constraints, organizations of all sizes are utilizing the ease-of-management and density of blade servers to support their critical HPC and enterprise applications. Because BladeRunner is easy to deploy, manage and service, it drastically reduces administrative costs. BladeRunner's built-in redundancy significantly improves system availability. "Although cluster and blade based solutions have captured more than a third of technical server market revenues, expanded adoption will depend in large part on vendors' ability to create products that appeal to new classes of technical users, " said Chris Willard, research vice president at IDC. "By offering environmentally friendly, preconfigured systems, Penguin Computing is addressing two of the major requirements reported by end users. First, environmental friendliness is directly related to system scalability; the lower the power and cooling requirements of a cluster node, the greater potential for scaling up a cluster before hitting environmental limits. Second, clustered systems are often difficult to configure, manage and maintain. Clusters that save on dollars per Gflops at the expense of increased system management costs are unacceptable to a large percentage of end-users." BladeRunner supports 12 dual-processor blades in a 4U chassis and can be configured in a 42U rack with 240 processors. Competing products require a 7U chassis for similar processing power. The integrated switch further reduces space and cabling requirements. BladeRunner's simple and efficient design includes powerful low-voltage Intel Xeon CPUs, which reduce the cost associated with cooling server farms. The entire unit, including a field-replaceable midplane, can be maintained without the need for special tools. BladeRunner includes integrated management software to monitor and control hardware components remotely.