An adaptive control of Web QoS based on online identification

Abstract

In this paper an adaptive control of Web QoS based on online identification is designed and implemented which dynamically adjust parameters of proportional-integral (PI) controller according to the changes of system model. In order to achieve admission control a session-based admission time-ratio feedback control mechanism is introduced which manages the amount of requests that the server can accept during the control period. The simulation results of Apache Web server compared with fixed PI controller demonstrate that the adaptive control based on online system identification is able to control the system resources of the server more effectually in case of overload and further improves the service quality.