Enhancing Cloud Performance with Clara and Nature Inspired Paradigm

Abstract

Cloud Computing is the structured IT model that pools services together as an infrastructure platform or software. Based on pay-as-needed, Cloud providers provide these to users. As the demand for these services or resources is increasing day by day, the resource- sharing may initiate a problem of availability. Hence, it is intricate to handle the load by the cloud providers at this rate. Subsequently, the load balancing concept comes into existence. Load balancing means to distribute the workload among all virtual machines (VMs) in a manner that distinctly each is fully utilized but not over-utilized. If anyone of the VM is over-utilized, the task is removed from that VM and given to an underutilized virtual machine (UVM). It includes two concepts, which are task scheduling and high resource utilization. Task scheduling and management of the resources are closely related to each other. In this context, clustering concepts use Nature-inspired Meta-heuristic Algorithm and Multi-Objective Task Scheduling. Load Balancing is an NP-Hard problem; hence, the Meta-heuristic Nature-inspired technique suits well to solve. newlineThe first part of the research discusses the meta-heuristic algorithms inspired by nature and the clustering algorithms. A study is also carried out in order to compare and contrast the clustering algorithms: K-Means Clustering and Clustering Large Applications (CLARA) Clustering. The two methods were tested to find lacuna in the current techniques. newlineThe second part of the research work includes the implementation of CLARA Clustering on multi-objective task scheduling using multiple Quality of Services (QoS) parameters and fault tolerance of cloudlets and virtual machines, for enhancement of the performance. The enhancement of performance is achieved by minimizing the results, including Execution time, Makespan time, Cost, and maximizing the throughput using Artificial Bee Colony (ABC) Optimization. newline