Using Automated Control Charts for the Runtime Evaluation of QoS Attributes


As modern software systems operate in a highly dynamic context, they have to adapt their behaviour in response to changes in their operational environment or/and requirements. Triggering adaptation depends on detecting quality of service (QoS) violations by comparing observed QoS values to predefined thresholds. These threshold-based adaptation approaches result in late adaptations as they wait until violations have occurred. This may lead to undesired consequences such as late response to critical events. In this paper we introduce a statistical approach CREQA - Control Charts for the Runtime Evaluation of QoS Attributes. This approach estimates at runtime capability of a system, and then it monitors and provides early detection of any changes in QoS values allowing timely intervention in order to prevent undesired consequences. We validated our approach using a series of experiments and response time datasets from real world web services.

In Proceedings of IEEE 13th International Symposium on High-Assurance Systems Engineering (HASE)
Ayman A. Amin
Associate Professor of Statistics