Benchmark simulator with dynamic environment for job scheduling in grid computing
Job scheduling algorithm has a significant influence on grid computing performance. Characteristics of jobs and resources to be used in evaluating the performance of the scheduling algorithm must reflect the dynamic nature of real grid environment.Static models of jobs and resources cannot be used t...
| Main Author: | |
|---|---|
| Format: | Conference or Workshop Item |
| Language: | English |
| Published: |
2014
|
| Subjects: | |
| Online Access: | https://repo.uum.edu.my/id/eprint/13091/1/Ku.pdf |
| Summary: | Job scheduling algorithm has a significant influence on grid computing performance. Characteristics of jobs and resources to be used in evaluating the performance of the scheduling algorithm must reflect the dynamic nature of real grid environment.Static models of jobs and resources cannot be used to generate jobs and resources in simulating the grid environment because of the dynamic nature of the grid.This paper presents a new graph representation of jobs and resources which is practical for hybrid metaheuristic model implementation such as ant colony optimization and genetic algorithm.A dynamic model that can generate jobs and resources similar to the jobs and resources in the real grid environment is also proposed.Jobs and resources may join in or drop out from the grid.Stochastic analysis is performed on the characteristics of jobs and resources.A simulator based on the dynamic expected time to compute, has been developed and can be used as a benchmark.The
simulator can generate jobs and resources with the characteristics of jobs and resources in the real grid environment.This will facilitates the evaluation of dynamic job scheduling algorithm. |
|---|