Please use this identifier to cite or link to this item:
|Title:||Energy Efficient Scheduling Approach for Grid|
Chana, Inderveer (Supervisor)
|Abstract:||Grid Computing computes the large-scale and complex problems in science, engi- neering and commerce by uniting the power of disseminated resources dynamically based on their availability, potentiality, makespan, cost and users Quality of Ser- vice (QoS) requirements. In Grid computing, complex and large problems are divided into small problems and distributed over the resources. Thus, grid com- puting increases the computational power through effi cient resource utilization. Grid Computing faces many challenges like resource management, security, energy-e ciency, Load balancing etc. In Grid Computing, it is very di cult to map the resources with tasks effi ciently due to fluctuation in users requirements and to achieve better performance within budget and time. The energy consump- tion of this large scale distributed system is also of important concern. Reducing energy consumption increases the execution time of a task on a processing ele- ment; however, the overall energy consumption may decrease but it decreases the performance as well. Thus simultaneous optimization of energy and performance in Grid Computing is a big challenge. This challenge can be addressed by effi cient management and scheduling of tasks and resources to reduce energy consumption without degrading the performance. In this thesis, energy and performance aware layered Grid architecture has been proposed for the implementation of Energy-effi cient and High Performance(EEHP) algorithm. This architecture contains three layers in which, the fi rst layer, a Grid Portal has been designed and presented for the submission of tasks, the second layer, a Provisioning Manager takes care of performance and energy factors by analyzing the submitted tasks and resources and the third layer, Scheduler maps the tasks to resources according to the proposed EEHP algorithm. This algorithm attempts to map more complex and dependent tasks to more energy and perfor- mance effi cient resources and thus minimizes the energy. Gridsim Toolkit has been used to validate the experimental results. These experimental results demonstrate that the proposed approach reduces energy more e efficiently as compared to the existing algorithms without degrading the performance.|
|Appears in Collections:||Masters Theses@CSED|
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.