EMC ScaleIO for High-Performance Computing

Challenges of High-Performance Computing

High-performance computing (HPC) environments are typical in a variety of industries today—oil and gas, financial services, and pharmaceuticals, to name just a few. With compute farms of often hundreds or thousands of nodes, HPC applications perform many concurrent multi‑threaded, multi‑client data‑access operations, resulting in massively parallel workloads that require cross‑node data sharing. Moreover, rates at which data is loaded typically fluctuate unpredictably.

HPC has a direct impact on time to market: the more compute nodes you have, the faster the jobs are completed, realizing a shorter time to market and to revenues. Therefore, every budget dollar not spent on compute power has a negative impact on profit. The demanding large-scale characteristics of HPC require traditional SAN storage and fabric that are typically complex and costly—at the expense of compute nodes and the bottom line.

EMC ScaleIO—the Ideal HPC Storage Solution

For HPC environments, ScaleIO provides an ideal alternative to SAN. ScaleIO provides capabilities similar to those of traditional SAN, along with the following benefits:

  • Performance and capacity increase linearly with the addition of compute nodes and disks.
  • ScaleIO scales out to hundreds and thousands of compute nodes.
  • Elasticity enables you to seamlessly add, move, and remove compute nodes on the fly during I/O operations.
  • Parallel architecture supports massive fluctuating concurrent I/O workloads across many compute nodes.
  • Self-healing functionality makes labor-intensive storage tasks a thing of the past.

With all these capabilities, ScaleIO cuts HPC storage costs by more than 80%, delivering a direct savings of over 28% on an organization’s total IT budget. This savings can now be allocated to the purchase of more compute nodes—reducing your time to market and increasing your bottom line.

Learn more about EMC ScaleIO for high-performance computing.