On the effectiveness of isolation-based anomaly detection inrncloud data centers

摘要

The high volume of monitoring information generated by large-scale cloud infrastructures posesrna challenge to the capacity of cloud providers in detecting anomalies in the infrastructure. Traditionalrnanomaly detection methods are resource-intensive and computationally complex forrntraining and/or detection, what is undesirable in very dynamic and large-scale environment suchrnas clouds. Isolation-based methods have the advantage of low complexity for training and detectionrnand are optimized for detecting failures. In this work, we explore the feasibility of IsolationrnForest, an isolation-based anomaly detection method, to detect anomalies in large-scale cloudrndata centers.Wepropose amethodtocodetime-series information asextra attributes that enablerntemporal anomaly detection and establish its feasibility to adapt to seasonality and trends in therntime-series and to be applied online and in real-time.