New method for scheduling heterogeneous multi-installment systems

Since the past decade, the use of parallel and distributed systems has become more common. In these systems, a huge chunk of data or computations is distributed among many systems in order to obtain better performance. Dividing data is one of the challenges in this type of systems. Divisible Load Theory (DLT) is a proposed method for scheduling data distribution in parallel or distributed systems. Many studies have been done in this field but only a few articles about distributing data in a heterogeneous multi-installment system can be found. In this paper, we present some closed-form formulas for the different steps of scheduling jobs in a heterogeneous multi-installment system (finding the proper number of processors, the proper number of installments, closed-form formula for scheduling internal installments and closed-form formula for scheduling last installment). Two different systems are studied: Computation-Based Systems and Communication-Based Systems. The results of our experiments show that both methods gave better performances than the previous methods (Hsu et al.'s method and Beaumont et al.'s method) and the Communication-Based method has a smaller response time than the Computation-Based method.