Efficient query monitoring using adaptive multiple key hashing

Abstract

Monitoring continual queries or subscriptions is to determine the subset of all queries or subscriptions whose predicates match a given event. Predicates contain not only equality but also non-equality clauses. Event matching is usually accomplished by first identifying a "small" candidate set of subscriptions for an event and then determining the matched subscriptions from the candidate set. Prior work has focused on using equality clauses to identify the candidate set. However, we found that completely ignoring non-equality clauses can result in a much larger candidate set. In this paper, we present and evaluate an adaptive multiple key hashing (AMKH) method to judiciously include an effective subset of non-equality clauses in candidate set identification. Each subscription is mapped to a data point in a multidimensional space based on its predicate clauses. AMKH is then used to maintain subscriptions and perform event matching. AMKH further provides a controlling mechanism to limit the hash range of a non-equality clause, hence reducing the size of the candidate set. Simulations are conducted to study the performance of AMKH. The results show that (1) a small number of non-equality clauses can be effectively included by AMKH and (2) the attributes whose overall non-equality predicate clauses are most selective should be chosen for inclusion by AMKH.