Design and Analysis of Adaptive Load Balancing Approach in Cloud Infrastructure

Abstract

In this study an Adaptive Load Balancing (ALB) approach is developed to effectively balance the load distributed across the cloud servers to minimize bandwidth and energy consumption on service provisioning. Cloud computing infrastructure has evolved as highly scalable services with massive computation power and storage capability with the resources being provided as service by the cloud environment and guarantees the Service Level Agreement (SLA). However, the needs of the subscribers have grown to an extent that there requires a big active platform for load balancing even if the resources are shared. Besides, the cloud computing paradigm also needs to optimally balance the load at the middle of the servers in order to avoid hotspot and improve resource utility. To perform energy conservation in cloud infrastructures, the use of chronological traffic data from data centers uses a service request prediction model. Collaborative provable data possession scheme adopt Homomorphic verifiable responses and hash index hierarchy but the drawback is that the match index structure are not matched properly with clustering model. Different level of power tariffs and requests made to the servers affect the decisions, where to serve the cluster needs. SLA Laws on privacy includes a factor that decides whether the loads can be moved in or out of a cluster, whereas they affect the overall energy consumption. ALB approach balances the load from every cluster group by minimizing the bandwidth and energy consumption. With repetitive query messaging, ALB collects the information about the current load of other group and then computes the average energy and bandwidth consumption of each group. The ALB Approach not only balances the energy consumption but also enhances the utilization of resources with minimal bandwidth usage. Extensive level of experimental studies is conducted to illustrate the efficiency and effectiveness of the proposed method. An experimental evaluation is accepted out to estimate the performance of the ALB approach with Virtual Machine (VM) energy-efficient cloud data centers. Performance metric for evaluation of ALB approach is measured in terms of energy consumption, bandwidth utilization rate, performance tradeoff and response time to service request, load balance factor and clustering efficiency.