A Novel Network Flow Watermarking Technology Combined With Error Correcting Code And Interleaved Code

Computer networks bring conveniences and benefits,but network attacks lead to serious problems of information security.In order to hide real identities,network attackers usually use various anonymous technologies,such as anonymous communication systems and stepping-stones.Therefore,tracing sources of network attacks has become an important topic in the field of network security.However,the traditional methods of IP-based tracing and passive traffic analysis have encountered great challenges,because of the interferences of anonymous technology.Therefore,Network Flow Watermarking(NFW)was proposed to solve the problem of traffic tracing in complex networks.NFW actively manipulates the invariant features of the target network flow to embed special marks,and then track the network flow by identifying these marks.NFW methods can be divided into three classes:the class based on interval differences,the class based on Direct Sequence Spread Spectrum(DSSS)and the hybrid class.These methods can trace attack sources,but their performances are easily impacted by various network interferences in complex networks.Typical interferences include chaff packets,packet dropped,packet repacketization,flow splitting and timing perturbation.However,existing NFW methods have some flaws such as the low recognition rate of multi-flow tracing,the lack of error correction,the difficulty to guarantee reliability,and the sensitivity of judgment threshold.In this thesis,we propose a novel NFW technology combined with error correcting code and interleaved code,which has wide applicability and supports existing NFW methods.This technology can improve the NFW methods,because it fully adopts the error control principle of error correcting code and utilizes the interleaved code to enhance the correction ability from burst errors.Such that,the improved NFW has the ability to correct errors automatically,ensures the reliability of NFW transmission and improves the success rate of multi-flow identification.Moreover,the novel NFW technology no longer relies on the threshold to detect NFW,and then improves the flexibility and resistance of network interferences.This thesis successfully combines seven representative NFW methods with Bose-Chaudhuri-Hocquenghem(BCH)error correcting code and interleaved code.We evaluate these improved NFW methods in a unified experimental platform by the performance metrics of error correction and reliability.In this platform,we build a proxy server to control five major factors of network interferences and apply different degrees of interferences into the target network flow.The results of comparative experiments present that the improvement NFW methods outperforms the existing ones because of the introduction of BCH code and interleaving code.Finally,we propose a general model of performance analysis according to the characters of network interferences.