Sizing considerations for Spotfire Statistics Services
Spotfire Statistics Services contains three services that can connect to your Spotfire Server to run complex statistical analyses using data functions.
- The Spotfire Enterprise Runtime for R - Server Edition (a/k/a the TERR service) is used to execute statistical computations written in Spotfire Enterprise Runtime for R (TERR) or in R.
- The Spotfire Service for Python is used to execute statistical computations written in Python.
- The Spotfire Service for R is used to execute statistical computations written in R.
General insights
- The services must have enough memory to handle the data sent for computation and manipulation.
- By default, the number of computational engines is one less than the number of logical processor cores. The number of processor cores needed will depend on whether jobs are long or short running, and the number of concurrent jobs triggered.
Hardware requirements
- Each service should have a minimum of two cores, but four is recommended.
- The cores can be split across more than one node or computer for high availability.
- The services can be scaled independently.
- More cores are needed for data rich and complex computations or to execute more jobs more frequently.