Load Balancing in Cloud Computing Using Support Vector Machine and Optimized Dynamic Task Scheduling

D. Radhika,M. Duraipandian

DOI: https://doi.org/10.1109/icrito51393.2021.9596289

2021-09-03

Abstract:Cloud computing enables the internet based computing to everywhere. In cloud computing the task scheduling and load balancing is the important issue that must be considered in data centers. In cloud platform service providers allocate the resources based on the available resources this will increase the QoS. The process of distributing workloads is a load balancing Technique. In this paper focusing the load balancing between the virtual machine. Load balancing are different types that is network load, memory capacity, storage capacity and CPU load. Load balancing supports failover, enabling scalability and decreases costs and maximizes the resource availability. This paper presents the detailed review of load balancing with machine learning technique such as naïve bayes, random forest, Decision Tree and Support Vector Machine. By these machine learning model, SVM algorithm significantly to classify the incoming request to reduce the processing time (makespan). Also provide the moth flame optimization technique for dynamic task scheduling to make better scheduling under different load condition. We have experimented and evaluated this algorithm in a cloud simulation environment; we obtain better outcome than some other well-known algorithms. The Simulation results demonstrate that the proposed technique has a better development in terms of reducing the makespan time and load balancing.