الفهرس | Only 14 pages are availabe for public view |
Abstract In chapter 1 a general introduction to the to the thesis topic and the motivation of the thesis topic then shows thesis contribution and brief outline of thesis content. Chapter 2 - E-learning with Virtualization and containers: a brief background about webservices then shows its advantage and benefits then moved to show how was the digital transformation has affected on the performance of eLearning systems as a web application. After that provided overviewed traditional virtualization and virtualization core components with focus on hypervisors, virtual machines architecture and show how this traditional virtualization technology is no longer suitable with new generation of application deployment in favor if new virtual technology called containers which proven high impact on deployed application and comparison between virtual machines and containers. Discuss how the cloud and its architecture have benefited from containers in deploying high performance applications with less hardware consumable and discuss the container orchestration platform Kubernetes which depended on in this research to deploy the web application. Chapter 3 – State of the Art: In this chapter the related work to this thesis is discussed in two parts. In part one, containers autoscaling issue. In part two, discussing the related work regarding the containers scheduling issue and how the metaheuristics optimization algorithms were used to enhance the resource scheduling optimization. Chapter 4 - Secure Based Predictive Autoscaling Model For containerized application: in this chapter started to show the steps followed to deploy the Kubernetes cluster with monitoring application Prometheus and visualization of monitored resources and then deploy web application were started to apply the enhancement methos of this research. Then discussed the contribution enhancement methods was followed to enhance the deployed containerized application. Machine learning module used to predict healthy hosts before deploying web application to decrease the fails possibilities of containers deployments. Clean the incoming traffic to containers web application to save the required resources to handle this fake workload, apply content caching to save the resources and internet bandwidth consumed. Predict application future workload and apply the autoscaling required before the sudden workload to avoid application bottleneck. Chapter 5 - Hybrid metaheuristic Multi-Objective Optimization Scheduling Method For containerized application: in this chapter briefly overviewed the container scheduling and the classification of Scheduling optimization algorithms then show how metaheuristics optimization algorithms was used to optimize the container scheduling and aiming to maximize the nodes resources utilizations and decrease costs. Chapter 6 – Concludes this study. It becomes clear that the proposed container autoscaling using Machine learning forecasting and using metaheuristic to optimize Kubernetes scheduling process have achieved high performance and was able to achieve high resource utilization and able to save the cost spent. |