Performance Measurement for Parallel and Distributed Programs: a Structured and Automatic Approach
IEEE Transactions on Software Engineering
Distributed Application Performance, Metrics and Management
Proceedings of the IFIP TC6/WG6.1 International Conference on Open Distributed Processing II
Correlating resource demand information with ARM data for application services
Proceedings of the 1st international workshop on Software and performance
Parameter estimation for performance models of distributed application systems
CASCON '95 Proceedings of the 1995 conference of the Centre for Advanced Studies on Collaborative research
Runtime metrics collection for middleware supported adaptation of mobile applications
Proceedings of the 5th workshop on Adaptive and reflective middleware (ARM '06)
| Hi-index | 0.00 |
Performance measurement systems and management applications are needed to help control and manage the performance behaviour of distributed applications. Measurement systems are responsible for efficiently collecting information about the applications and making it available for use by management applications. Efficiency requires the performance information for each application process to be aggregated on a periodic basis, before it is made accessible to management applications. As a result, errors in performance estimates for application metrics can arise when deducing performance measures for the processes and the distributed application as a whole. A loose upper bound for the worst case error is given, and a simulation study is conducted to better quantify the error for an example system. We consider the impact of this error on several different types of management applications. Lastly, a technique is presented that can be used to decrease the error without sacrificing efficiency.