BeeGFS and Western Digital

The BeeGFS parallel file system was tested using IOR and the mdtest benchmarks on six client servers. Ultrastar DC SN840 NVMe SSDs were partitioned into 100GB and 1.89TB namespaces for metadata and object storage, respectively. OpenZFS was used to create metadata and storage targets, offering robust data integrity and fast RAID schemes. A striped zpool was created on each server, with the 1.89TB namespaces configured as ZFS RAID- Z2 (6+2) pools. The zpools were formatted as ZFS file systems. BeeGFS Sequential Read/Write Performance The BeeGFS file system, configured with ZFS pools "mraid001" and "sraid00[1-4]", achieved 1.6GB/s read and 3.1GB/s write throughput with a single IOR thread. When all 36 cores on a single client were used, read and write throughput reached 11GB/s, fully saturating the 100GbE network link. BeeGFS on OpenFlex Data24 Deployment Topology

BeeGFS on Western Digital's OpenFlex Data24 Storage with xiRAID is a user-friendly, easy-to-manage solution for IT/HPC. It offers an open composable infrastructure environment and disaggregated compute on the NVMe-oF storage platform. The performance of BeeGFS on OpenFlex Data24 systems shows excellent benchmark results for storage data streaming with both RAID and Non-RAID configurations. The solution enables data protection and high availability, accelerating storage hardware systems to new levels of performance. This self-service, composable infrastructure solution helps customers transition to a more flexible, variable cost model.