Mark Claypool, Robert Kinicki, and Craig Wills
Today's home wireless networks concurrently support a plethora of user applications with diverse connectivity and Quality of Service (QoS) requirements. A solution to performance issues is a `smart' Access Point (AP) that automatically improves performance in an interoperable, easy-to-use fashion that a user expects. However, prior to designing such an AP, techniques that identify and classify applications based on wireless network traffic characteristics are needed to enable the AP to respond accordingly to diverse QoS demands. Rather than classify applications, our approach concentrates on the nature of traffic due to specific applications and devices. By characterizing flow signatures in terms of the nature of the traffic, we aim to develop a better and more effective matching strategy between the resultant signatures and the potential treatments that need to be applied to wireless traffic flows.