Research On Resource Allocation For Anti-traceable Access Networks

To trace and control the trend of online public opinion and even the international situation,government agencies,emergency management departments and other legitimate institutions and organizations in the real world need to gather and analyze data from domestic and foreign news media,specific websites and popular online social platforms in a covert way to avoid being traced by particular individuals or organization.The essence of anti-traceable access is to hide information browsing track and search intention through advanced technologies to prevent visitors from being traced.A secure access system can provide a safe,controllable and anti-IP tracking access environment.The main means of trackback is to reconstruct the access path of the visitors.Therefore,constructing the access path which is difficult for the attacker to reconstruct will be one of the important means to resist trackback attacks.It is of great significance to build a secure access environment based on link scheduling and resource allocation.In this thesis,focus on secure access system,combined with the needs of the project "Research on targeted convert collection and anti-traceable Access for Internet Data[BH2018-CF03-1]" with China Academy of Electronics and Information Technology of CETC,we conducts a comprehensive study on link scheduling and resource allocation.The main contributions are described as follows:(1)To tradeoff the access delay and the guaranteed security level,a dual objective joint optimization algorithm of link scheduling is proposed to hide users' access trace.The resource allocation problem of secure transmission in multi-hop VPN network is discussed to obtain the relationship between resource allocation and security level guarantee.Taking the type of data links,transmission constraints and queue model into account,a mathematical model of the relationship between link hops and security level requirements is established.Furthermore,we propose a novel resource allocation and link scheduling algorithm with guaranteed security level.The resource allocation with the minimum delay can guarantee the security level.Finally,the objective function is transformed into linear integer programming by piecewise linearization,and the approximate solution is achieved.The approximation ratio between the solution of mixed integer linear programming and the solution of the objective function is (1+?).(2)Focus on marginal utility caused by long path,i.e.,the utility of anti-traceable performance gradually decreases with the increase of the length of path.To measure the impact of the path length on anti-traceable performance,an anti-traceable resource allocation algorithm based on reinforcement learning is proposed by introducing marginal utility function.For a VPN network with no dispatching center,the autonomous path of nodes is constructed,and the Q-learning algorithm of delay optimization is introduced.Through the recursive analysis of delay,the state transition model of multi-hop VPN network is established.Based on the theory of marginal utility,a group of marginal utility factors with different security levels and delay are selected,and the strict strategy and easing strategy are explored to tradeoff security requirement and delay.Experimental results and performance analysis verify the impact of two utility factors on resource allocation.