In this video from ISC 2019 in Frankfurt, Dale Brantley from Panasas describes how the company delivers a frustration-free HPC storage environment.
Panasas delivers HPC data storage solutions that support industry and research innovation around the world. Whether it’s building the next Dreamliner, winning a Formula One race, creating mind-bending visual effects, curing disease, or modeling climate change, the world’s leading companies trust Panasas to support their most innovative HPC projects.
At ISC, Panasas showcased its new ActiveStor Ultra storage system with PanFS 8, the industry’s only plug-and-play portable parallel file system. ActiveStor Ultra supports the most innovative high-performance computing and AI projects with limitless performance scaling, management simplicity, and consistently high performance regardless of workload complexity.
The Panasas ActiveStor solution uses a scale-out architecture that grows storage capacity, DRAM caching, and network bandwidth incrementally and linearly as you add more ActiveStor enclosures. It delivers data from storage nodes in parallel to the application, multiplying the bandwidth an application can achieve to a single file, not just aggregate bandwidth. And data flows directly from our storage nodes to the application without any hops through intermediate servers or even extra network links.
ActiveStor Highlights:
- Simple. The Panasas scale-out ActiveStor solution is a single entity you manage from one graphical user interface (GUI) or command-line interface (CLI), no matter how many ActiveStor enclosures you integrate into it. It takes only a couple hours to install the ActiveStor solution. Just rack it up, plug it in, power it on, answer a few questions, and it’s ready for anything you can throw at it. Even in the largest Panasas deployments, all data resides within a single namespace, with a single management GUI and CLI, delivering data at very high reliability and availability. It is possible to quickly add more ActiveStor enclosures; and each added unit will immediately contribute more capacity and performance. The Panasas ActiveStor solution will automatically rebalance capacity across the ActiveStor enclosures as you add them or if they become unbalanced; automatically reconstruct the full levels of erasure-coded data protection for all files in the event of any failures; and continuously scan all files in the background to scrub out any latent issues.
- Flexible. The Panasas architecture scales data and metadata independently and is purpose-built for adaptability and flexibility to handle a wide range of use cases. The Panasas ActiveStor solution automatically adapts to dynamically changing workloads and increasing demands. Its scale-out nature inherently spreads the workload, reducing the impact of hot spots, as well as simply growing capacity and performance. ActiveStor product families offer flexible configuration options for each node to meet specific workflow needs, now and in the future. The solution supports mixing generations of ActiveStor products within a single namespace.
- Cost-Effective. The number of people it takes to keep a storage system performing at its peak also affects its value. Part of our focus on HPC storage for the past two decades has been on reducing complexity, automating failure management, and supporting wide and widely mixed workloads without requiring tuning or retuning. It takes only part-time attention from a single person to manage the Panasas ActiveStor solution, no matter how large or high performance it is.
Dale Brantly leads the global Panasas systems engineering group responsible for architecting and delivering ActiveStor storage solutions to customers worldwide. Dale has more than 30 years of experience in defining engineering requirements and achieving architectural and quality excellence for complex storage solutions deployed in a large high performance computing and enterprise environments. Prior to joining Panasas, Dale managed global solutions development at Silicon Graphics (SGI) and ran SGI’s Information Lifecycle Management (ILM) product line. Other appointments include field and customer site technical positions at Cray Research, Control Data Corporation, and ETA Systems.
Sign up for our insideHPC Newsletter
