Abstract
This study addresses a duplication based scheduling algorithm called Effective Scheduling based on Task Duplication (ESTD) for grid computing environment. Duplications are made based on task dependencies. The algorithm ensures beneficial duplications and avoids unnecessary duplications. Idle time slots between task execution times are effectively used. The algorithm aims to avoid the communication contention, which will happen when there is frequent transportation of large sets of data. The performance of the algorithm is scaled by comparing it with the algorithms of its kind. The results show minimized make span and effective resource utilization with balanced loads across resources in grid.
Highlights
Grid computing distinguishes itself from other parallel and distributed systems through its unique features which may include heterogeneity of computing resources and their dynamic participation, varied administrative domains of resources and networks, dynamic accessibility of resources by the users and so on
Resource Management System (RMS) sets the permissions for grid users to use the grid resources and tracks the resource utilization, which would be used for billing purpose The Grid Resource Broker System (GRBS) is an interface between grid service providers and grid users; it helps to use the available grid resources
When considering the compared algorithms, the result consistencies on makespan is remarkable for Effective Scheduling based on Task Duplication (ESTD)
Summary
Grid computing distinguishes itself from other parallel and distributed systems through its unique features which may include heterogeneity of computing resources and their dynamic participation, varied administrative domains of resources and networks, dynamic accessibility of resources by the users and so on. Grids are formed out of dedicated resources that form a parallel computer; use the computing power of personally owned computers that are available on the internet. Later category of resources let their computing cycles to be used whenever they are connected over internet and necessarily their CPU cycles are found idle. These resources are identified as non dedicated resources. Grid computing emphasizes on proper resource utilization and shortest makespan in executing the tasks. In such environments, the resources cannot be kept idle as it could lead to maximum completion time of tasks. An application is a group of tasks which may have interdependencies among them
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have