Architecture, Implementation, and Deployment of a High Performance, High Capacity Resilient Mass Storage Server (RMSS)

The Naval Oceanographic Office (NAVOCEANO) High Performance Computing (HPC) Major Shared Resource Center (MSRC) recently reengineered the existing mass storage system serving its high-performance compute platforms. The purpose was to provide significantly improved file system availability, to refresh the technology and the architectural design, and to position the MSRC to incorporate emerging technologies such as Storage Area Networks (SANs). The resultant configuration utilizes two SUN Enterprise 10000s (E10K) with 3 TB of switched Fibre Channel Disk Arrays and the latest generation of tape devices. The theoretical peak system capacity is in excess of 2 Terabytes (TB) per day, with management of up to 1 PetaBytes (PB) of storage and an aggregate external network throughput of 220 Megabytes (MB) per second. This paper discusses the technologies and considerations used in the design of the MSRC Resilient Mass Storage Server (RMSS), its architecture, implementation and integration, deployment and transition from the existing CRAY Data Migration Facility (DMF).