Abstract
Network Traffic Classification is very essential for Internet service providers (ISPs) and helping them in defining and adapting the priorities. A lot of decisions and actions are taken depending on the traffic classification. That is for both commercial and technical issues. In this paper, different approaches were developed to classify the internet traffic data from different point of views, namely, Traffic Types, URL Categories and Visited websites. The classification was done for each of these types depending on traffic amount and the number of accessing counts. These approaches were applied on a data collected and obtained from the Deep Packet Inspection (DPI) tool installed in the Libyan International Telecom Company (LITC) network, which is representing the only international gateway to Libya through different ports. In order to reduce the international traffic, LITC installed cache servers on its network four years ago. The used data in this work were taken before and after installing these cache servers, to show the changed happened to the network and how cache severs affecting the amount and the nature of the international traffic. The result shows that a large portion of the internet data passing through the Libya gateways is for social network, streaming and search engines, especially for Facebook, YouTube, and Google websites. And the impact of installing cache servers was dominant in these categories