A Detection Technology Based On Network Behavioral Characteristics Of Trojan

With the expansion of the network, the new network services emerging to meet the people fast, flexible access to information needs, but a variety of network problems follow. Network Trojan deterioration of the normal network environment, society as a whole had a negative impact, the Internet "content security" problem is getting people’s attention. Quickly and efficiently detect and prevent network further Trojans will inevitably become a major challenge currently facing and problems, network-based behavioral characteristics Trojan detection technology is based on a network flow detection depth Trojan detection technology, which to some extent has been put into practical complements the characteristics of the fingerprint detection technology, to meet the actual needs.Topics proposed matrix method based on evolutionary behavior for network Trojans feature detection characteristics distinguish characteristics of flow and non-flow, thereby efficiently and accurately identify a variety of Trojans. Currently, the industry application concerns mostly Trojan network data encryption has been adopted, confusion and other technical means, resulting in no significant packet content features, this experimentally derived characteristics traffic classification accuracy rate can reach91.5%, the results show that the method with a more high specific network Trojan detection accuracy.Finally, this paper introduces the research group developed a prototype Trojan detection systems, the system is divided into fingerprint detection, and network behavior feature detection in two parts, in order to detect the fingerprint-based, ensuring fine-grained detection, detection of network behavior to compensate for deep packet inspection some weaknesses:For deep packet inspection possible misidentification, network behavior feature detection for identification and prompt re-testing to achieve error correction purposes, and through the formation of the final decision is the recognition result.This research can effectively improve the specific network Trojan detection accuracy and efficiency of the Internet to help maintain good health of the environment, thus creating a harmonious network society.