Enhancing the Traffic Privacy with Leakage Detection and Optimization

摘要: With recent experimental works, it is found that there a information leakage in the packetized flow of data due to user activity and traffic content. We aim at understanding how complex when leaked by packet features namely length, direction times. Technique call this type feature called masking. Here, we define security model find ideal target masking which removes any leakage. Further, investigate tradeoff between privacy protection making cost, required amount overhead realization complexity feasibility. Major findings are 1). Masking packets network. achieves similar values with padding only case fragmentation allowed 2) Optimized statistical attains moderately better than simple fixed pattern does, while still leaking correlation can be exploited.