When a Reactor (an instance of a Decision Service) processes rules, it accesses the data resident in “working memory”. Working memory is populated by any of the following methods:
1. The payload of the Corticon Request: In the installation options described in Choose the deployment architecture, Option 1 and 2 express the payload as an XML or JSON document. In Option 3, the payload consists of a reference to Java business objects. Regardless of form, the data is inserted into working memory when the client’s request (invocation) is received. When running a Studio Test, the Studio itself is acting as the client, and it inserts the data from the Input Ruletest into working memory.
2. The results of rules. During rule processing, some rules may create new data, modify existing data, or even delete data. These updates are maintained in working memory until the Rulesheet completes execution.
3. An external relational database. If, during the course of rule execution, some data is required which is not already present in working memory, then the Reactor asks Corticon Server to query and retrieve it from an external database. For database access to occur, Corticon Server or Studio Test must be configured correctly and the Vocabulary must be mapped to the database schema.
