Metrics for monitoring a hierarchical service-based system

The dynamic nature of service infrastructure causes the Quality of Service (QoS) of Service-Based Systems (SBS) to be unpredictable during execution. This necessitates monitoring and analysis tools which collect, measure and analyse metrics that characterise performance of a SBS from diverse facades in order to recognize elements that contribute to anomalies. To have a clear insight into the working of a business process, performance metrics should be defined and measured from multiple facets of the SBS. Existing approaches which attempt to manage performance of a SBS focus on metrics of the business process, services or infrastructure resources. There is a lacuna in studies which define and assimilate metrics from all levels in order to gain a holistic view of the business process performance. In this study we define metrics at design-time to manage performance and obtain metrics from all the levels of an SBS during execution using Complex Event Process (CEP) paradigm. The empirical results demonstrate that our approach has negligible effect on the performance of the SBS.