ACCURATE AND REAL TIME METHOD FOR NETWORK PACKET CLASSIFICATION

Abstract:- Evolution of Internet has been accompanied by the development of a range of network applications which introduce many interesting, challenging and tricky problem to the network traffic monitoring. The increasing numbers of applications cause network traffic become harder to classify. Classifying internet traffic by looking at headers of packet and the transport protocol port number is no more reliable and accurate technique for the reason in today's computer world, running well-known applications on not-so-well-known ports or used others protocol as wrappers is a widespread scenario. Meanwhile, others examine packet payload in order to identify degree of reliability and accuracy of traffic classification. Yet, due to some restriction such as security, privacy and legal issues, examining the packet payload could be restrained in such cases. However, accurate and real-time methods that can dependably recognize the application that generated a flow still need to be developed without involved in neither human interventions nor offline classification to achieve efficient and effective packet verification. Hence, a common problem with many protocol analyzers is the inability to accurately identify and consequently decode a protocol that runs over a non-default port number. In order to optimize network investment; all packets should be clearly identified; nevertheless, there still no such tool in the market today is capable to analyze all network protocols. In our work, we put on our effort to advance the network monitoring system via an accurate, reliable and robust real time application. In view o