Journals Proceedings

Abstract

Using of parallel and distributed system has become more common. Dividing data is one of the big challenges in this type of systems. Divisible Load Theory (DLT) is one of the proposed methods for scheduling data in parallel and distributed systems. Recent research has applied divisible load theory in real-time scheduling and has been introduced as an alternative for multiprocessor scheduling.There are two type of scheduling algorithm in real-time divisible load theory(RT-DLT) which is known as single-round and multi-round algorithm. Most studies in this field are about distributing data in single-round algorithm. Unfortunately, multi-round algorithms are difficult to analyze and have received only limited attention in real-time concept. In this paper, we will determine the minimum number of processors needed to complete the job by its deadline in multi-round algorithm. The two algorithms are compared on linear programming based formulation and result show that multi-round algorithm can provide a significant improvement on minimum number of processor needed in comparison with single-round algorithm.