Research On Cross-domain Collaborative Tracking Technology Based On Network Flow Watermarking

With the rapid development of Internet,the network scale extends continuously and the applications spread widely.Meanwhile,the security situation is not optimistic enough.Network security incidents occur frequently,which not only disrupt the normal order of the network but also cause huge economic lost.Intrusion traceback technology is used to discover the true location of the attackers in the network automatically.This technique can both discover the attack packets transmission path in a network and provide the necessary decision basis for intrusion detection event handling.Intrusion traceback produces a powerful deterrent to the attackers,thus reducing the frequency of invasion behavior effectively.However,the existing network agent,NAT technology,IP tunnel and anonymous communication system provide convenience to the attackers to hide their real IP.In addition,attackers usually exploit stepping-stones spanning multiple autonomous system,which increases the complexity of the intrusion traceback technology.In this paper,the general situation of network security was analyzed.Firstly,the design specifications for a cross-domain collaborative intrusion tracking framework were proposed.The functional components and implementation process of this framework were introduced in detail.Such a framework can prevent the leaks of sensitive information among autonomous systems and keep the network operation boundaries.Secondly,a Histogram Specified Interval Centroid-based Watermarking(HSICBW)scheme was proposed.The HSICBW can highly improve the resistance of the Interval Centroid-based Watermarking to Multi-flow Attacks,thus it reduces the time and space complexity of watermark detecting.To improve the applicability of existing watermarking technology,this paper proposed an equalization preprocessing method for Interval Centroid-based Watermarking that can both enhance the stability of the watermark carrier and the robustness of network watermarking.Compared with the existing technologies,these two methods proposed in this paper achieve better effectiveness of confidentiality,robustness and adaptive ability.