Prediction Based Load Balancing in Cloud Computing Using Conservative Q-Learning Algorithm

Cloud computing is a striking expertise trend that provides Computing assets as a service experiencing a revolution for the IT industry and academics researchers. The potential of emergent cloud computing technology is efficiently hooked by the primary requisite such as resource management. The furthermost vital aspect of asset controlling techniques in Cloud environment depends on scheduling and Load Balancing techniques. Load Balancing strategy is attracting and generating considerable interest in order to utilize resources, thereby increasing the enactment of the cloud datacenter. The energy consumption in the datacenter is customarily due to improper utilization of resources in terms of overloading the Servers or sometimes due to idle Servers. Load is flourishing in the recent era; the Internet based Computing proposals shared resources such as hardware, software, and information on the demand basis. Cloud Computing carry amendments, and the revolution of the Information Technology industries emerged with its popularization and applications. Balancing is one of the best solutions for efficient utilization of resources and is extensively considered to be the most important method to decrease energy utilization. The prime aim of the exploration work is to scrutinize various Load Balancing techniques and propose an energy-aware Load Balancing strategy for ideal utilization of the assets in cloud Computing environment. The proposed Conservative Q-learning algorithm (CQA) for maintaining efficient equilibrium between the work load among virtual machines and optimally lessens the energy ingesting through the Load Balancing algorithms proposed by us.