Authors:
Abu Asaduzzaman
1
;
Manira Rani
1
and
Darryl Koivisto
2
Affiliations:
1
Florida Atlantic University, United States
;
2
Architecture Modeling Group, Mirabilis Design, Inc., United States
Keyword(s):
Distributed and parallel system, scheduling algorithm, load balancing, performance analysis, VisualSim.
Related
Ontology
Subjects/Areas/Topics:
Distributed and Mobile Software Systems
;
Energy and Economy
;
Load Balancing in Smart Grids
;
Parallel and High Performance Computing
;
Smart Grids
;
Software Engineering
Abstract:
The concurrency in a distributed and parallel system can be used to improve the performance of that system by properly distributing the tasks among the processors. However, the advantage of parallelism may be offset by the increased complexity of load balancing techniques. Scheduling is proven to be an effective technique for load balancing in any distributed and parallel system. Studies indicate that for application-specific systems static scheduling may be the potential choice due to its simplicity. In this paper, we analyze the performance of load balancing by static scheduling for distributed and parallel systems. Using VisualSim, we develop a simulation program that models a system with three processors working simultaneously on a single problem. We obtain the response time and completion time for different scheduling algorithms and task groups. Simulation results show that load balancing by scheduling has significant impact on the performance of distributed and parallel systems.