Object Caches Overview and Best Practices

An object cache is a recently used object definition stored in the memory of Desktop and Intelligence Server. You can create object caches for both application objects (project, report, document, etc) and schema objects (facts, attributes,etc).

The example below illustrates how object cache works:

• A user opens the Report Editor.
• The collection of objects displayed in the Report Editor make up the report’s definition. If no object cache for the report exists in the Desktop or the Intelligence Server memory, the object request is sent to the metadata.
• The report object definition is retrieved from the metadata and displayed to the user in the Report Editor
• An object cache is created in memory of the Intelligence Server and the Desktop machines. If the same user requests the same object, the object cache on the Desktop machine satisfies that request. If a different user requests the same object, the object cache in Intelligence Server satisfies that request.

Object Caching Best Practices

Allocate enough memory in Intelligence Server for the storage of object caches. Tuning this setting to allocate memory greater than the 100 MB default is recommended when dealing with complex schemas with a large number of objects. Object cache calculation is based on the metadata size. Object caches should range from 20-50% of the metadata size. A metadata size of 1 GB should have an object cache of 500 MB, which corresponds to 50% of the metadata size.