HPC STORAGE
.JPG)
Clustered Storage:
The main problem statement for HPC based storage is the requirement to support high performance, efficient, scalable I/O. This means multiple GB/s of parallel bandwidth distributed to tens, hundreds or even thousands of compute nodes. In lay terms, as the number of computed nodes increases in a HPC cluster so must the I/O capabilities of storage system. Traditional Network File System (NFS), Network Attached Storage (NAS), or Storage Area Networks can bottle neck and not scale their I/O without literally adding whole systems. So employing parallel storage architecture that scales along with supporting ever increasing green initiatives is a must.
Contact