Comparative study of machine learning algorithms for anomaly detection in Cloud infrastructure

Abstract

Cloud is one of the emerging technologies in the field of computer science and is extremely popular because of its use of elastic resources to provide optimized, cost-effective and on-demand services. As technology started to grow in scale and complexity, the need for automated anomaly detection and monitoring system has become important. Inappropriate exploitation of Cloud resources can often lead to faults like crashing of VMs, decreased efficiency of cloud system etc. thereby leading to violations of the Service Level Agreement (SLA). These faults are often preceded by anomalies in the behavior of the VMs. Hence, the anomalies can be used as indicators of faults which potentially violate the SLAs. We have created a system that will monitor the VMs, detect anomalies and warn the system administrator before any problem escalates. We present in this paper a comparative study of various machine learning algorithms used for detecting anomalies in cloud.