SwiftStack has launched a new product for big data analytics, machine learning hybrid cloud and multi-cloud environments.
The platform was designed for data-driven workloads that use frameworks and applications such as Hadoop, Spark, Presto, TensorFlow and Hive. According to SwiftStack, it enables users to create a big data analytics or artificial intelligence-machine learning data pipeline in a memory-first architecture where storage and compute are decoupled and can scale on-demand as data load or performance needs grow.
This new product is powered by SwiftStack's object-based cloud storage and 1space multi-cloud data management. It also uses Alluxio, a data orchestration layer that sits between computer frameworks and storage. SwiftStack claims this eliminates common challenges such as a lack of enterprise-ready multi-cloud workflows with appropriate data management tools, insufficient throughput, or insufficient API compatibility to support descriptive, interactive and predictive analytical workload.
To deploy and scale this product, users can add standard on-premises server hardware and account capacity in AWS or Google Cloud Platform. Applications and users get universal access to a single data namespace, and data is placed across the system by user-defined policies.
SwiftStack launched 1space in 2018, which creates a single storage namespace. This enables data migration while running an application, which SearchStorage called a major contribution to the open source community.
Additionally, SearchStorage predicted multi-cloud storage as a hot trend for 2019, including SwiftStack's 1space product, and specifically named the analytical element as the next step for multi-cloud data management.
According to SwiftStack, its products are most used for functions such as artificial intelligence or machine learning, analytics, scientific research, active archive and managing data across multiple clouds.