摘要

In this paper, a novel quality of service (QoS) prediction framework is proposed to accommodate different applications with various QoS requirements. The purpose of the proposed structure is to estimate, given the QoS requirements of different applications, the number of them can be satisfied according to the predefined QoS level and the number of them have to be rejected or accepted under lower level. Moreover, this structure is able to control the network behavior when the new application with certain QoS requirements requesting for the connection. The framework consists of four modules. QoS Requirements Monitoring detects the changes in the supplied QoS. QoS Prediction performs at real-time to estimate whether the new QoS can be adapted to the current QoS. QoS Violation is dedicated to detect the violation. And a module of QoS Estimation is used to make the decision whether the hopeless new traffic is refused or accepted with acceptable QoS requirements. Therefore, this structure is able to control network behavior when the surrounding environment changes. Numerical and simulation results obtained suggest that the proposed predictive scheme is a promising approach for real time QoS prediction.