ANOMALY DETECTION IN THE SERVICES PROVIDED BY MULTI CLOUD ARCHITECTURES: A SURVEY

Mahendra Kumar Ahirwar, Manish Kumar Ahirwar, Uday Chourasia

Abstract: An Anomaly is abnormal activity or deviation from the normal behavior. Anomaly detection is the process of removing these abnormal or anomalous behaviors from the data or services. The services delivered to users by cloud service providers must have normal behavior. To provide services to users in the proper and normal form, anomaly detection becomes important and interested area for research work. For anomaly detection so many techniques are developed and these techniques are broadly divided into three categories: - statistical, data mining based and machine learning based anomaly detection technique. Anomaly detection techniques are used to detect and discard anomalies from the data or services. In this survey paper we provide overview of some anomaly detection techniques which are discovered recently for the tracing data. In the anomaly detection models anomalies are detected by comparing the tracing data with the actual data. On the basis of comparison deviations in the traced data or services are identified and they are considered as anomaly.

Keywords: Anomaly, services, tracing data, anomaly detection techniques and cloud service provider.

DOI: https://doi.org/10.15623/ijret.2014.0309030