Spotfire® Server and Environment - Installation and Administration

Sizing considerations for Spotfire Server

The Spotfire Server contains a central application server which is made up of all the installed Spotfire servers working in a cluster. The server also includes the internal services Library Services, User Services, Audit Services Deployment Services and Information Services.

Service Description
Library Services The centralized storage for Spotfire analysis files and metadata.
User Services Provides user authentication and role-based authorization.
Audit Services The centralized collection of action logs (authentication, content modification and access, and permission changes).
Deployment Services Provides automatic delivery of client product upgrades and patches.
Information Services A centralized point of data access and metadata management against enterprise relational data sources. Data managed with Information Services streams down to the in-memory engines with optional caching to disk. The approximate RAM usage for the service is 100MB per concurrent data connection.

The service is capable of supporting about 60 to 80 concurrent data connections with the suggested hardware. Typical data retrieval speeds are greater than 2.5 million cells per second per connection, assuming sufficient database and network bandwidth.

Hardware requirements

Hardware sizing needs to accommodate the central application server and these five services. In general, for the application server, plan for a minimum of two Spotfire Server servers and a separate database server for the Spotfire database. The following guidelines can help you determine the best strategy for minimizing cost and maximizing efficiency.

  • Spotfire Server should have a minimum of two processor cores. Four is recommended.
  • The four cores could be split between two clustered Spotfire Servers with two cores each on different computers for high availability.
  • Additional hardware requirements depends mainly on the Information Services load, specifically:
    • The data sources and JDBC drivers used.
    • The number of simultaneously executing information links.
    • The number and type of joins in each information link.
    • The number and type of transformations in each information link.