Graph-based rule and transaction execution in a parallel active object-oriented knowledge base management system
Ramamohanrao S. Jawadi, Stanley Y. W. Su · University of Florida Digital Collections (University of Florida) · 1994
The need for user-defined execution orders (or control structures) for rules is well recognized by researchers of active database systems. Priority-based approaches have been used to specify and enforce control structures among event-condition-action (ECA) rules. However, due to the fact that fixed priorities are assigned to ECA rules independent of different contexts in which they may be triggered, the existing approaches are not able to allow rules to be executed following different control structures when they are triggered by different events. In a priority-based approach, the control specification is mixed with the rule specification which makes the control difficult to understand and modify. More flexible and expressive control structures are needed for rules in advanced database applications such as CAD/CAM, CIM and Flexible Manufacturing Systems. The rule control mechanism must also be incorporated in the object-oriented data model so that control structures among rules can be modeled uniformly. Since rules can be activated by operations in database transactions and their executions need to be incorporated in the transaction framework, a powerful transaction model is needed to deal with complex control structures among rules. Unfortunately, the variants of the nested transaction model used in several existing active systems are not expressive enough to handle rules with complex control structures in a uniform fashion. In this thesis, we introduce the concept of rule graphs in which control structures among rules are clearly separated from rules. Using rule graphs, the same set of rules may follow different control structures when they are triggered by different events. We uniformly extend an object-oriented knowledge model, which models rules as objects, to capture control relationships among rules uniformly. A graph-based transaction model is used to model graph-based control structures among database operations and rules in a uniform fashion. The proposed object-oriented knowledge model, rule and transaction modeling and execution techniques have been implemented and verified on a shared-nothing multiprocessor computer nCUBE2 which exploits the parallel execution properties of independent rules and DB operations with a graph-based control structure.