期刊名称：International Journal of Computer Networks (IJCN)
电子版ISSN：1985-4129
出版年度：2012
卷号：4
期号：3
页码：53-71
出版社：Computer Science Journals
摘要：The network have in the world highly congested channels and topology which was dynamically created with high risk. In this we need flow classifier to find the packet movement in the network. In this paper we have to be developed and evaluated TCP/UDP/FTP/ICMP based on payload information and port numbers and number of flags in the packet for highly flow of packets in the network. The primary motivations of this paper all the valuable protocols are used legally to process find out the end user by using payload packet inspection, and also used evaluations hypothesis testing approach. The effective use of tamper resistant flow classifier has used in one network contexts domain and developed in a different Berkeley and Cambridge, the classification and accuracy was easily found through the packet inspection by using different flags in the packets. While supervised classifier training specific to the new domain results in much better classification accuracy, we also formed a new approach to determine malicious packet and find a packet flow classifier and send correct packet to destination address.
