Grid computing is the preferred high performance data processing architecture in many companies, but storage response issues and added hardware increase costs and complexity. While most initial grid deployments start out their life-cycle with fairly well-balanced performance, performance bottlenecks increase as the demands on the grid grow and greater numbers of faster execution nodes are added.

Read this white paper to learn about an approach that addresses this problem and has the following features:

• 48 TB raw capacity per server
• A data mover component that moves data between the Lustre file system and long term storage significantly faster than traditional network transfer methods
• A separate management network with connections to all cluster nodes
• High-performance SunFire™ servers, the Lustre scalable cluster file system, and high-speed/low-latency InfiniBand interconnects, fully tested and integrated into one system