Panasas, Inc.: PanFS on ActiveStor Ultra Built for Price/Performance - Engineered for Manageability
TimeWednesday, June 24th8:09pm - 8:23pm
DescriptionIn this video as fast paced as their product, Panasas presents an overview of their new, competitively priced, COTS-based, ActiveStor Ultra HPC storage appliance that is 2X faster than Lustre, BeeGFS and GPFS, without their notorious complexity and brittleness.
Panasas’ unique Adaptive Small Component Acceleration (ASCA) architecture is what makes Panasas so fast. Small files are stored on low-latency flash SSDs and large files are stored on low-cost, high-capacity, and high-bandwidth HDDs. Metadata is stored on ultra-low latency NVMe SSDs and an NVDIMM intent-log stages both data & metadata operations that’s 2,560X lower latency than the NVMe media used in competitive systems.
Panasas ASCA provides a single “all-hot” tier for instantaneous access to files no matter where they are in the data lifecycle – with no latency or waiting to re-migrate data to a flash tier as all other parallel file systems have to do. ASCA is what makes Panasas perform so well on mixed workloads and why Panasas delivers such predictable and scalable performance – no matter what workload, or changing workload, is thrown at their systems.
Hyperion Research surveyed HPC sites around the world and found they experienced frequent failures and lost productivity. In contrast, Panasas mission-critical HPC storage provides all the performance and price-performance HPC organizations are after, without complexity and downtime. HPC organizations no longer need to trade off performance and price-performance for simplicity, uptime, and great support. With the new Panasas ActiveStor Ultra, HPC organizations can have it all.
Not yet registered? Event registration is available here .
Director of Product Management