International Journal of Recent Technology and Engineering (IJRTE) ISSN: 2277-3878, Volume-8 Issue-3, September 2019 577 Published By: Blue Eyes Intelligence Engineering & Sciences Publication Retrieval Number: B2041078219/19©BEIESP DOI:10.35940/ijrte.B2041.098319 Classifying Internet Traffic using An Efficient Classifier Haitham A.Jamil, Hind G. Abdelrahim, Bushra M Ali, Azza O. Awad Abstract— The new development in the architecture of Internet has increased internet traffic. The introduction of Peer to Peer (P2P) applications are affecting the performance of traditional internet applications. Network optimization is used to monitor and manage the internet traffic and improve the performance of internet applications. The existing optimizations methods are not able to provide better management for networks. Machine learning (ML) is one of the familiar techniques to handle the internet traffic. It is used to identify and reduce the traffic. The lack of relevant datasets have reduced the performance of ML techniques in classification of internet traffic. The aim of the research is to develop a hybrid classifier to classify the internet traffic data and mitigate the traffic. The proposed method is deployed in the classification of traffic traces of University Technology Malaysia. The method has produced an accuracy of 98.3% with less computation time. Keywords— Classifier, Internet traffic, P2P, Internet traffic Mitigation, Machine Learning, Neural Network I. INTRODUCTION P2P application is one of the familiar developments in internet architecture[1]. Traffic data classification is used to reduce the internet traffic. The classified data will inspect the network and try to consume less bandwidth[2]. Managing network traffic is the major challenge in distributed network. Existing researches are not commendable in the area of P2P computing. The differences in the properties of traffic data will affect the performance of network. ML methods are used to develop an automated application to deal complex data. Dataset and feature selection are the important factors of classifiers. The ability of classifiers are dependent on these factors. ML is a set of technique and data analytical methods to improve the performance of a automated task. ML algorithms are widely used in real – time applications[3]. Revised Manuscript Received on September 15, 2019 Haitham A. Jamil, University of Elimam Elmahdi, Kosti, White Nile, Sudan haithamjamil@mahdi.edu.sd Bushra M Ali, University Technology Malaysia, Johor Bahru, Malaysia bushra0912115@gmail.com Hind G. Abdelrahim, University Technology Malaysia, Johor Bahru, Malaysia hindjamil33@gmail.com Azza O. Awad,University Technology Malaysia, Johor Bahru, Malaysia azzaawadelkareem@gmail.com Support Vector Machine, C5.0, and Artificial Neural Network (ANN) are some of the ML techniques used for the classification of P2P network. SVM is a supervised method that uses mapping function for the classification[4]. The mapping function is used to classify the data according to the labels. The classified model represents the data that was used in the training phase[5]. C5.0 is a familiar technique used as an alternative for SVM. It is based on Decision tree algorithm. The concept of estimation of entropy is used to take decsions[6][7]. The features of dataset will be used for the derivation of patterns and matched with a target class[8][9][10]. J48 algorithm is used in the selection of features by evaluating a node of the tree. It is a flexible method and easy to combine with other ML techniques[11][12][13]. Most of the complex ML models were built with NN and produced an optimal solution. It is basically a time consuming method[15][16]. A proper training is required for NN to produce good results. The objective of the research is to provide a efficient Artificial NN (ANN) for the classification of P2P internet traffic. ANN is combined with J48 to improve the computation time of the classifier[17][18]. The proposed model can be used to identify the emergence of new P2P traffic application in distributed networks. The state of the art classifiers are compared with the proposed method in terms of accuracy and computation cost[18]. The structure of the paper is arranged as follows: section two will provide information about the existing literature on network traffic classification. Section three will give details about the methodology of the research. Section four will discuss the experimental setup of the research. Section five will provide results and analysis and finally, the paper will be concluded in section six. II. REVIEW OF LITERATURE Internet classification is used to provide better quality of service is managing network traffic. A classifier will cluster the internet traffic data into different groups. The new improvements such as dynamic port numbers an packet payloads are complex and difficult to classify using older classification technique.ML algorithms are need to be employed to classify complex communication data. Pramitha P et.al. ,(2018)[1] have compared ML algorithms for classification of internet traffic data. The authors have employed Naïve Bayes, Random Forest(RF), Decision Tree(DT), and Multi layer perceptron algorithms for the classification of internet traffic. The results have shown that RF and DT have better accuracy than other classifiers.