IaaS Signature Change Detection with Performance Noise

Abstract

We propose a novel framework to detect changes in the performance behavior of an IaaS service. The proposed framework leverages the concept of the IaaS signature to represent an IaaS service's long-term performance behavior. A new type of performance signature called categorical IaaS signature is introduced to represent the performance behavior more accurately. A novel performance noise model is proposed to accurately identify IaaS performance noise and accurate changes in the performance behavior of an IaaS service. A set of experiments based on real-world datasets is carried out to evaluate the effectiveness of the proposed framework.

15 pages, accepted and to be published in 19th International Conference on Service Oriented Computing. arXiv admin note: text overlap with arXiv:2007.11705