Fuzzy-Smith Control For Qos-Adaptive Notification Service

Yuying Wang,Xingshe Zhou
DOI: https://doi.org/10.1007/978-3-540-73549-6_111
2007-01-01
Abstract:In this paper, a QoS-Adaptive Notification Service (QANS) scheme is proposed to prevent system overload, in which the Fuzzy-Smith control approach is applied. Different with traditional admission control, QANS adaptively adjusts the QoS level of task to control the service frequency of service provider by real-time monitoring the event channel. As a result, the system load is controlled and the performance of notification service is guaranteed. QANS is independent of the mathematical model of event channel and provides a non-linear control action, which eliminate the control error caused by inaccurate system model and linear assumption. Experimental results show that the proposed QANS scheme outperforms traditional event channel controllers in terms of system load control, response delay, and task miss-rate.
What problem does this paper attempt to address?