Adaptive Hybrid Auto Scaler for Resources Scaling in Cloud Computing Environment

Abstract

Auto-Scaling is one of the main features of cloud computing. Auto-scaling facilitates the customer to scale up and scale down the resources as per the requirements. This feature of the cloud improves the flexibility and elasticity of the cloud resources. Auto-scaling consists of complex cloud operations that support the fluctuating demands that are coming from customers. For implementing this feature of the cloud, we need the prediction of the future workload and according to the incoming rate of the workload the present work manages the resources in advance; however, it leads to the problem of under-utilization of resources. And for solving the problem of under-utilization of resources, design of a Hybrid Auto Scaler (HAS) to entertain the autonomic-provisioning of resources. HAS predicts the future behavior of the incoming requests using the Time series method (Auto-Regression of order one) and computes the necessary resources in advance using the analytical method. HAS framework also takes care of the efficiency of load balancing.