Probabilistic Optimized Kernel Naive Bayesian Cloud Resource Allocation System

Cloud service providers offer diverse utilities to the consumer on their heterogenous requirements which may fit on different levels of performance measurements such as costs, quality of service, and accuracy, etc. Cloud is enriched with unlimited resources along with many advanced features such as scalability, robustness which increased the cloud demand in recent years. The selection of appropriate resources is a challenging task that can minimize the cost, and maximize the resource utilization and consumer experience. We have proposed the optimized kernel naive bayes cloud service selection model (OKNB) which works on the concept of maximum probability. The cloud resource bundle with maximum probability is chosen as a predicted cloud resource. Our proposed model achieves 88.76%, 88.45%, and 93.65% accuracy on response time, CPU utilization, and memory utilization models which is 3.15 and 8.04% higher than the state of the art models on response time and memory utilization models respectively. The proposed model yields 16.83 and 29.175 s lower waiting time in standard deviation and mean value compare to the GMP-SVM model.