This thesis proposes and implements an efficient network traffic classification method based on a new vital random forest for high dimensional data. Network traffic engineering is one of the most important technologies that have witnessed a rapid growth in the revolution of worldwide technologies. Network traffic classification has added considerable interest as an important network engineering tool for network security, network design, as well as network monitoring and management. It can introduce different services such as identifying the applications which are most consuming for network resources, it represents the core part of automated intrusion detection systems, it helps to detect anomaly applications and it helps to know the widely-used applications for the intention of offering new products. On the other hand, several challenges faced by network engineers on their course to classify traffic. The most common of which are increasing application types and the huge size of data traffics. Therefore, many researchers have been competing in literature to introduce an efficient method for traffic classification. The efficiency is dependent on important factors such as classification accuracy, memory consumption and processing time. This thesis presents a Vital Random Forest (VRF) as efficient network traffic classification which is a one package that introduces a new features-selection technique, data inputs reduction and a new build model for original random forest method to classify network traffic for huge datasets. VRF aims to reduce processing time, increase classification accuracy and decrease memory consumption.
