WekaIO and SDSC are collaborating on a range of technology projects, including high performance I/O research. SDSC's Advanced Technology Lab (ATL) will leverage a research computing cluster and WekaIO Matrix storage software to evaluate the technology for accelerating diverse research workloads on a fast flash layer. Wall-clock application performance, as well as I/O performance, will be measured and benchmarked on scientific workloads running on traditional parallel file system architectures and those with "burst buffer" augmentation.
Unlike traditional storage solutions, WekaIO's Matrix eliminates performance bottlenecks by aggregating local SSDs inside the servers into one logical pool, which is then presented to host applications. MatrixFS, a fully distributed, massively parallel file system, delivers unprecedented performance to latency sensitive applications. A fully integrated tiering layer transparently offloads cold data to any S3 or Swift Cloud object store, scaling to exabytes of data in a single namespace for easy management.
"SDSC supports hundreds of multidisciplinary research programmes spanning a wide variety of domains, from astrophysics and bioinformatics to environmental sciences and health IT", stated Ron Hawkins, SDSC's director of Industry Relations. "This engagement with WekaIO has the potential to allow emerging high-performance I/O technology to be leveraged more broadly across the research community and at the same time directly benefit our own efforts in data-intensive computing innovation."
Accelerating discovery through the elimination of storage bottlenecks is of high interest to the general research and genomics communities, and this collaboration will permit the evaluation of next generation I/O technology for its suitability to support computational methods leading to new scientific and scholarly discoveries.
"SDSC is a recognized leader in HPC, scientific research, and data analytics, providing infrastructure and expertise to the national research community", stated Barbara Murphy, vice president of Marketing at WekaIO. "Legacy parallel file systems were developed prior to the emergence of Flash in the data centre, whereas WekaIO's Matrix file system (MatrixFS) has been written to run natively on SSD and NVMe devices, to deliver unrivaled performance to data intensive applications. We are delighted to partner with the talented people at SDSC on this important initiative."