Real-time traffic classification for unified communication networks

The key feature of Unified Communications (UC) lies in its ability to provide the service that the customer wants, when he wants and regardless of the device being used. At the heart of the UC architecture, which spans different layers from application layer to network, comes the Quality of Service (QoS) framework that is essential to provide adequate service to different communication session's call participates. The heterogeneity of access networks and diversity of a UC session's participants make it necessary to classify incoming traffic based on its type and the QoS requirements of the participating parties. The existence of different traffic classifications for different existing access networks raises the question of how to classify incoming traffic when it belongs to a diverse set of access networks and devices. In this paper, we investigate different classification methods currently adopted in different access networks and industry standards. We introduce a traffic classification method that should satisfy the requirements of realtime traffic that traverses most of the existing access networks. The classification method has been tested using both First In First Out (FIFO) and Weighted Fair Queuing (WFQ) schedulers to evaluate different classes performance in a shared environment with diverse QoS requirements. The results obtained showed the feasibility of our proposed classification.