A Cognitive QoS Method Based on Parameter Sensitivity

Different applications in the network have different sensitivity for the certain QoS parameters. The existing mechanisms cannot modify the packet loss policy exactly according to the needs of QoS parameters. Therefore, the network will not maximize its overall efficiency. This paper proposes a novel cognitive approach for QoS. It classifies the applications by various combinations of the QoS parameters. And it uses a cognitive layer to modify the loss probability of the packet. Simulation results prove that this approach can reduce the account of useless packets sent when the congestion occurs in the network. The overall efficiency of the useful data transmission can be improved as well. The method can be used by Internet of Things to improve its message processing and sensor data fusion to reduce energy consumption.