Modern business needs for data management have resulted in a surge in the fields of data science, machine learning, and big data platforms. These business needs (e.g. marketing, advertising, and hiring) are essential to both end-products and business critical operations and require large amounts of computational resources. Improper allocation and scheduling of these resources can cost companies millions of dollars per year. Sub-optimal resource allocation is often time consuming to diagnose and requires experienced engineers many weeks to solve. While modern infrastructure-as-a-service platforms, including Amazon Web Services and Google Cloud provide easy interfaces for scaling computational needs, efficient allocation and scheduling of these resources are still necessary to efficiently utilize them. Thorium fills these data management gaps.
About
Sandia's Thorium job scheduler is an architecture-agnostic job scheduler that allows users to run data processing jobs with zero deployment files. System users such as data engineers and business development teams can submit computational tasks to perform and Thorium provisions and allocates resources on a compute environment for the task to run. Thorium then ensures that the job is created, scheduled, executed, and completed successfully. Thorium can scale the compute environment from a laptop to a large compute cluster.
Key Benefits
Efficient provisioning and allocation of resources on a compute environment Does not require complex coding or engineering to scale eliminates the need for large engineering support staff Deploys on top of any Kubernetes cluster Ensures that the job is created, scheduled, executed, and completed successfully
Applications
Job scheduler Handling of large datasets and processing infrastructure