Publication

Abstract : Resources provisioning is one of the key challenging task in cloud computing environment due to dynamic provisioning of resources to the applications. In cloud environment, resources should be allocated dynamically to the application, according to the workload behavior pattern of the application. Variations, in resources provisioning leads wastage of cost and energy. Additionally, it causes Quality of Service (QoS) dropping and Service Level Agreements (SLA) violations. So, the resources which was allocated to the applications, should be close to the current workload of the application. To achieve flexible elasticity mechanism in cloud environment, determination of future demand of resources to the application is the crucial step before the resources allocation. Thus, prediction of the resources to the application in dissimilar manners like number of virtual machine and physical machines, workload and number of requests etc., is an essential step. Based on the prediction results, enough resources can be allocated to the applications in the right time, while avoiding SLA violation and QoS dropping. This paper presented challenges, characteristics and importance of several prediction schemes to enhance workload management system in cloud environment.