A query-based horizontal fragmentation approach for disjunctive deductive databases

A disjunctive deductive database (DDDB) is more general than the deductive database model since the former captures more information than the latter. Distributing a DDDB facilitates the splitting of queries of the database into subqueries to enhance query processing since the answers to each subquery are computed only against those portions of the DDDB relevant to the subquery and can be generated in parallel. Evaluating these subqueries thus improves the response time of the system and reduces communication cost involved in evaluating the entire query. Existing literature on DDDBs does not include a well-designed horizontal fragmentation approach. In this paper, we propose a horizontal fragmentation approach for any given DDDB D based on the set of common user queries S of D from which a fragment is produced for each query Q in S. All the rules and ground facts in D required to answer Q are collected into a fragment F, and it is sufficient to evaluate Q against F alone to compute all answers to Q. Hence, the need to process fragments generated by the other queries of D is eliminated, and time and cost of unnecessary processing is saved