A rule termination analysis method for active rule management systems

Abstract

Active rule management system is an advanced facility which enhances traditional database functionality by providing not only database-related triggers but also application-related rule management. In the active rule management systems, operations are executed automatically by executing a set of pre-defined rules in response to events, such as database events, application events, etc. Active rules typically consist of event, condition, and action which are generally referred to as ECA rules. During the active rule executions, it is usual that rules fire other rule. In some undesirable situations, rules ate fired infinitely forming cycles. This phenomenon is said to be a non-termination of rules and should be avoided . The goal of rule termination analysis is to detect rule subsets that may potentially lead to infinite firing of rules. In this thesis, we present an efficient rule termination analysis method that can be used to verify the termination of fired rules in active rule based systems. We also present a development of a graphical user interface which allows the rule developer to be able to easily identify undesirable design of event-rule firing chain. With the identified cyclic rule chains. the rule developer can modify the specification of event and rules, and then, apply the rule termination analysis method again. This iterative process should be performed until termination analysis guarantees that the rule set will never manifest a non-terminating rule behavior. Many event-trigger-rule processing systems adopt composite events to give more expressiveness in defining application rules. Rule firing by composite events makes the rule termination much more complicated and thus hard to analyze . Most of existing termination analysis methods do not handle this problem properly. Some method reports the capability of detecting cyclic rule firing caused by composite events. However, they still have difficulty to tell genuine cycles from spurious ones. The rule...