An index structure supporting rule activation in pervasive applications

Rule mechanism has been widely used in many areas, such as databases, artificial intelligent and pervasive computing. In a rule mechanism, rule activation decides which rules are activated, when the rules are activated, and which tuples can be generated through the activation. Rule activation determines the efficiency of rule mechanism. In this article, we define the semantic constraints, constant constraint and variable constraint, of the rule according to the semantics of Datalog rules. Based on the constraints, we propose an index structure, named Yield index, to support the rule activation effectively. Yield index consists of the data index and semantic index, and records the complete information of a rule, including the matching relationship among the tuples of different relations in rule body. The index integrates tuple insertion and rule activation to directly determine whether the matching tuples of new inserted tuple exist. Due to this character, we perform effective rule activation only, avoiding ineffective rule activation that cannot generate new tuples, so that the efficiency of rule activation is improved. The article describes the structure of Yield index, the construction and maintenance algorithms, and the rule activation algorithm based on Yield index. The experimental results show that Yield index has better performance and improves activation efficiency of one order of magnitude, comparing with other index structures. In addition, we also discuss the possible extensions of Yield index in other applications.