Using magpie for request extraction and workload modelling
OSDI'04 Proceedings of the 6th conference on Symposium on Opearting Systems Design & Implementation - Volume 6
ACM Computing Surveys (CSUR)
Self-adaptive software system monitoring for performance anomaly localization
Proceedings of the 8th ACM international conference on Autonomic computing
Detecting application-level failures in component-based Internet services
IEEE Transactions on Neural Networks
MSEPT'12 Proceedings of the 2012 international conference on Multicore Software Engineering, Performance, and Tools
| Hi-index | 0.00 |
To allow architectural self-adaptation at runtime, software systems require continuous monitoring capabilities to observe and to reflect on their innate runtime behavior. For software systems in productive operation, the monitoring overhead has to be kept deliberately small. By consequence, a trade-off between the monitoring coverage and the resulting effort for data collection and analysis is necessary. In this paper, we present a framework that allows for autonomic on-demand adaptation of the monitoring coverage at runtime. We employ our self-adaptive monitoring approach to investigate performance anomalies in component-based software systems. The approach is based on goal-oriented monitoring rules specified with the OCL. The continuous evaluation of the monitoring rules enables to zoom into the internal realization of a component, if it behaves anomalous. Our tool support is based on the Eclipse Modeling Project and the Kieker monitoring framework. We provide evaluations of the monitoring overhead and the anomaly rating procedure using the JPetStore reference application as a Java EE-based test system.