The Panasas ActiveStor 5000 parallel storage cluster is the premier Panasas product family that combines an innovative blade architecture with the advanced ActiveScale Operating Environment that includes the PanFS parallel file system. The ActiveStor 5000 is designed for HPC data centers that are running both interactive applications on workstations and batch-oriented HPC applications on Linux clusters. The range of typical applications include oil and gas seismic processing and interpretation, electronic circuit design and chip simulation, automotive and aeronautical design and simulation, and animation creation and rendering. The requirements for these applications typically involve the widest range of workloads, from small to medium-sized files with more random access and higher I/O demands, to sequential access for very large files.

To support a unified storage infrastructure, the ActiveStor 5000 offers full high-availability providing maximum uptime across the HPC data center. The ActiveStor 5000 includes three important enhancements to manage these requirements: ActiveGuard, ActiveImage and StorageBlade XC.

Panasas ActiveGuard High Availability Software contains the enhanced high availability feature set that provides:

• Seamless failover of DirectorBlade modules to ensure continuous performance and access to all volumes
• Automated failover of the storage network to maintain continuous access to all volumes

Panasas ActiveImage Snapshot delivers easy, instant backup and instantaneous restore through an efficient snapshot implementation that only records differences since the last snapshot.

The Panasas StorageBlade XC provides 2 GB of high-speed cache memory to each StorageBlade module, or up to 20 GB per 10 TB shelf. This larger cache dramatically improves the already fast random I/O performance for small and medium-sized files.

HPC sites often have separate computing infrastructures using the same datasets for multiple purposes. The Panasas ActiveStor 5000 parallel storage cluster eliminates this inefficiency and reduces management complexity by facilitating a unified pool of storage under a single namespace that can serve both batch and interactive application needs in clustered computing environments.