Apprenticeship Learning Based Load Balancing Technique for Cloud Environment

Abstract

Cloud Computing is accessing and handling data and documents from the internet rather than from any individual computer hard drive. The issues faced by cloud computing are security, privacy, vendor lock-in, server downtime, network connectivity, dependency, vulnerability to attacks, load balancing, etc. Load balancing in cloud computing is one of the important issues as huge amount of load need to be efficiently distributed among the servers. The existing approaches to address load balancing issue are throttled technology, active clustering, central policy for virtual machine, round robin technology, max-min min-min, fuzzy monitoring, honeybee foraging behavior, reinforcement learning, etc. The primary drawbacks of above-mentioned approaches towards load balancing are lowered throughput, high migration rate, overloading and under-loading of resources. This paper proposes a novel architecture which applies apprenticeship learning for load balancing in the cloud. Its performance is found to be good with respect to parameters like response time, accuracy, learning rate and speed.