The most performant, dense, scalable, and flexible data storage platform for Spark

Modern applications that use Spark to analyze structured and unstructured data need to be future proofed. Reduce storage over provisioning, accommodate storage growth, consolidate applications and their data, speed time to results, and reduce TCO by disaggregating Spark data storage with the Pavilion HyperParallel Data PlatformTM. It is the industry’s first NVMe oF storage array. It scales to over a petabyte
of data, provides ultra-high performance, ultra-low latency, linear scalability, and uses NVMe SSDs.

A DAS-based Spark infrastructure results in 2-3X overprovisioning and islands of storage—while using software-defined storage takes CPU resources away from Spark, impacting time to results and increasing TCO. Pavilion changes all that with the Pavilion HyperParallel Data Platform.

Download the full NVMe-oF Storage for Spark solution brief to learn more.