It uses unsupervised algorithms along with an adaptive learning approach to automatically let classes of traffic emerge. To obtain a high-performance statistical feature. Select Classifiers allow that proper class labels can be easily discovered. Provide optimal classification accuracy and time Consuming. Accurate traffic classification is of fundamental importance to numerous other network activities from security monitoring. This is mainly related to the fact that port numbers are not embedded in a metric space. Thus ingenuity is required to smartly include them. In our work we engineer an iterative procedure to identify clusters of flows in which the srvPort information is used to filter elements in each cluster.