Automatic Traffic Classification Using Machine Learning Algorithm for Policy-Based Routing in UMTS-WLAN Interworking

摘要

The future mobile terminal will be dependent on the multiple wireless access technology simultaneously for accessing Internet to offer best Internet connectivity to the user. But providing such interworking among wireless heterogeneous networks and routing the selected traffic to particular wireless interface is a key challenge. Currently, existing algorithms are simple and proprietary, and there is no support to route the specific application traffic automatically. The proposed decision algorithm finds the optimal network by combining fuzzy logic system with multiple-attribute decision-making and uses na?ve Bayes classifier to classify the application traffic to route into appropriate interface to reduce the service cost. The performance analysis shows that the proposed algorithm efficiently uses the network resources by maintaining active connection simultaneously with 3G and Wi-Fi. It routes 71.99 % of application traffic using Wi-Fi network and 28.008 % of application traffic using UMTS network to reduce the service cost and to reduce network load on the cellular operator.