| The growing importance of Web traffic on the Internet makes it important that we have an accurate understanding of this traffic source in order to plan and provision. This thesis presents three key steps in the process of understanding process of the Web traffic: measurement, characterization, and modeling. We measured real Web traffic in the Georgia Tech campus network more than 20 different times over the course of two years. Web traffic is characterized in two protocol layers of HTTP and TCP in which the performance of Web is greatly affected. In the Web characterization, we decompose Web traffic into primary and secondary components and attempt to understand the implications of these components. Based upon knowledge learned in the characterization, a Web traffic model is constructed to represent the behavior of Web traffic in a simplified form. The resulting model produces a pattern of traffic, for use in network simulations, which is statistically similar to the pattern of traffic on a real network. The model was validated using a number of metrics. The study of Web traffic modeling enabled us to adapt a Medium Access Control (MAC) protocol for an Internet access via satellite. Our Web traffic model was used to evaluate performance of the MAC protocol after verifying the accuracy of the model. |
