VAST Data

VAST Data is a developer of a data platform designed for workload-intensive applications. The company works to develop data storage centers that do not use hard drives in order to increase the performance of data writing and access speeds. This includes the company's Universal Storage system, which works to offer flash media performance with features that offer a cost equivalent to hard drive-based archive systems. The platform also offers data management solutions for businesses, including system integration, technical services, service sales, and product sales.

VAST Data was founded in 2016 by CEO Renen Hallak, Shachar Fienblit, and Jeff Denworth, and is headquartered in New York City. Customers of VAST have included Lawrence Livermore National Laboratory, Jumptrading, Dug, Katapult, Verizon, National Institutes of Health, NASA, and the U.S. Department of Defense.

VAST Data developed its approach to building and deploying storage aimed at combining all-flash performance with archive economics at an exabyte scale. Universal Storage, the name of VAST Data's approach, combines these new storage technologies with a new class of storage algorithms designed to move beyond traditional hard drives by eliminating cost barriers to flash storage.

Universal Storage is offered to users as a cloud-native containerized storage architecture that eliminates storage tiering, and it is intended to allow users to analyze quickly and respond to real-time situations, especially in national security situations, while improving infrastructure performance for AI and HPC environments. The solutions for Universal Storage include use in life sciences, deep learning, animation and visual effects, quantitative trading, media and broadcast, and for the public sector, including for national security purposes.

Universal Storage is built on low-cost flash memory and storage class memory with new algorithmic concepts to offer fast storage access and fast networking over the system. This includes NVMe-over-Fabrics enabling commodity data center networks to transform into storage fabrics capable of scaling by use case. This is combined with QLC flash, which offers a new generation of high-density flash that enables VAST Data's concept and provides NVMe flash performance for applications. This is further combined with shared persistent memory, which uses a non-volatile storage media that makes it possible to field QLC in enterprise storage environments.

The benefits offered in this structure are that it removes cache coherency challenges with performance scaling linearly; it removes the concern around batteries or UPSs; there is no need for rebuilds during server failure due to the statelessness of VAST Data Universal storage, with failover being instantaneous; and it offers containerized architecture for autonomous and adaptive cluster scaling.

In VAST Data's update to Universal Storage in version 4.3, the company added metadata cataloging, security features, containers, and data movement to its operating system. In part, by request from its clients, VAST Data added data tiering with service plans with upper and lower bandwidth limits and IOPS per user or bucket, allowing users to distinguish between higher- and lower-performance workloads. Metadata cataloging was added as VAST Catalog, which allows user-defined context tagging of data in a tabular format to offer users insight into what is stored without needing third-party tools. And VAST Data added security features, including support for the Key Vault API, Thales CipherTrust, and IBM Key Protect cloud-based key management services while extending zero-trust work with support for TLS-based access for in-flight encryption.