Compute groups have these characteristics:
- With no service interruption, they can be started to accommodate bigger workloads and stopped when no longer needed (to lower costs).
- You can create them for specific departments, applications, or projects, and then segregate processing, monitoring, reporting, and billing by department, application, or project.
- You can associate each department with its own compute group, which includes one or more compute clusters. Associating each department with its own compute group lets you isolate each department's work and costs.
- They are ideal for periodic, predictable, special-purpose tasks, such as month-end processing, exploration, and projects.
For more information, see Managing Compute Resources.