Behavioral Analysis of Resource Allocation Systems in Cloud Infrastructure

This paper presents behavioral analysis of resource allocation systems in cloud infrastructure. This analysis is performed on the basis of these systems monitoring and machine learning. To do it, the system creates a model of normal system behavior using learning set. The constructed model is used as a reference. Recognition of an abnormal behavior of the system is based on the comparing of the normal behavior with the current one. The results represent the beginning of the research. In the continuation, the issues of anomaly detection based on virtual machine (instance) resource utilization metrics will be analyzed. Also, an algorithm for determining anomalies will be developed based on a comprehensive analysis of the CPU, RAM, HDD resources usage. For enhancing of the anomalies detection accuracy it is planned to increase the period of learning.