Resource Allocation for Cloud based E-Learning Systems

Abstract

Cloud computing has become one of the most promising IT solutions and business trends in recent years. As the need for cloud is increasing and the requirements are getting more diverse and complex, the need to find an effective allocation strategy is increasing. In such a virtualized shared environment the resource requirements of applications keep changing from time to time. And in such a scenario, static allocation fails to solve the purpose. It leads to the inefficient utilization of resources and might affect the quality of the application. In this paper the allocation of resources on a university cloud is discussed. Cloud computing proves to be a powerful platform for higher education. Although the adoption of cloud on the university level has various advantages, a lot of other factors like dynamically changing resource requirements, etc. need to be taken into account. Here, a dynamic allocation mechanism is presented by modeling the resource requirements using a normal distribution based on which an algorithm to solve the resource allocation problem is proposed. The algorithm ensures that a particular virtual machine is never overloaded. It also tries to minimize the difference between the resources demanded and allocated so as to maintain the quality of an application running on a virtual machine.

Keywords: Dynamic Resource Allocation, Load Balancer, Global Scheduler, Local Scheduler, Normal Distribution

View PDF