This document summarizes a student's research on using pattern matching to classify internet traffic. The student developed online and offline tools to collect data sets from different networks. A k-means clustering approach is used to model traffic types, where models are trained and new traffic is identified by calculating distances to the models. The performance of the approach is evaluated based on training data size, number of models, test data size, and features used. The research concludes the approach can easily identify traffic models and provide high bandwidth internet access for users.