Development of scheduling algorithm for cloud computing /
Cloud computing is one of the highly used type of distributed computing systems. One of the main benefit of cloud computing is the elasticity that gives on-demand services. Therefore, users can easily access services anytime and anywhere. In addition, cloud computing has been widely used for its eno...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
Kuala Lumpur :
Kulliyyah of Engineering, International Islamic University Malaysia,
2017
|
| Subjects: | |
| Online Access: | Click here to view 1st 24 pages of the thesis. Members can view fulltext at the specified PCs in the library. |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Cloud computing is one of the highly used type of distributed computing systems. One of the main benefit of cloud computing is the elasticity that gives on-demand services. Therefore, users can easily access services anytime and anywhere. In addition, cloud computing has been widely used for its enormous benefits and its ability to cope with large scale data such as workflows and big data applications. On the other hand, data are increasing dramatically by the use and easy access of Internet. This reduces the performance of cloud due to the huge data flow. Thus, the issue of scheduling rises, since a lack of appropriate scheduling methods will result to the failure of the potential benefits of the cloud systems. In order to optimize these resources, efficient scheduling algorithm are required. As a result, many scheduling algorithms, static or heuristics, are proposed by researchers to address big data challenges. In this study, an analysis of different scheduling algorithms used in computing systems and in cloud computing in particular was carried out. Since big data algorithms are multi-objectives in nature, these algorithms are not efficient enough because they are either single objective or they produce non-optimal solutions. Most of the time, these algorithms are not able to handle large number of tasks and the performance criteria is violated. In this dissertation, additionally, an architecture based on lambda concept for big data processing is introduced. Moreover, a meta-heuristic algorithm has been proposed. The proposed algorithm possesses the capabilities of resources optimization in a range finite of time. It is a meta-heuristic algorithm Particle Swarm Optimization based (PSO) that considers two objectives of scheduling namely meeting the deadlines and minimizing the cost. To evaluate the proposed algorithm, an extension of CloudSim simulator that support big data based applications is used. The results generated show that the proposed PSO-based algorithm schedules the tasks and performs more than 90% when compared to FCFS in term of optimizing the resources and the cost was minimized two times than FCFS. |
|---|---|
| Physical Description: | xiii, 102 leaves : illustrations ; 30cm. |
| Bibliography: | Includes bibliographical references (leaves 87-94). |