Comparison of Two Parallel Processing Methods for a Distributed Deductive Database