Multitasking without comprimise: a virtual machine evolution
OOPSLA '01 Proceedings of the 16th ACM SIGPLAN conference on Object-oriented programming, systems, languages, and applications
The Vision of Autonomic Computing
Computer
Introduction: Service-oriented computing
Communications of the ACM - Service-oriented computing
Effective Java (2nd Edition) (The Java Series)
Effective Java (2nd Edition) (The Java Series)
Program transformations for light-weight CPU accounting and control in the Java virtual machine
Higher-Order and Symbolic Computation
Classification of Component Vulnerabilities in Java Service Oriented Programming (SOP) Platforms
CBSE '08 Proceedings of the 11th International Symposium on Component-Based Software Engineering
A Practical Approach for Finding Stale References in a Dynamic Service Platform
CBSE '08 Proceedings of the 11th International Symposium on Component-Based Software Engineering
A Method for the Resource Monitoring of OSGi-based Software Components
SEAA '08 Proceedings of the 2008 34th Euromicro Conference Software Engineering and Advanced Applications
Towards a Service-Oriented Component Model for Autonomic Management
SCC '11 Proceedings of the 2011 IEEE International Conference on Services Computing
A self-healing component sandbox for untrustworthy third party code execution
CBSE'10 Proceedings of the 13th international conference on Component-Based Software Engineering
A-OSGi: a framework to support the construction of autonomic OSGi-based applications
International Journal of Autonomous and Adaptive Communications Systems
Hi-index | 0.00 |
In-production performance monitoring is required for dynamic and modular systems open to third-party applications such as the OSGi-based smart home box that home actors envision today. Existing approaches are not suitable for inproduction monitoring as they generally induce a strong persistent overhead. This paper proposes a progressive and localized monitoring system that is able to dynamically activate/deactivate and tune the accuracy of monitoring mechanisms depending on detected performance issues. In particular, it proposes to build a proxy-aware service registry in order to inject proxies on-the-fly without stopping bundles and it advocates the use of localized sampling. Our evaluation shows that the overhead of our monitoring system is only 2% when idle and comparable with traditional systems when completely active(around 20%).