Research On Data Encrypted Analysis Technology Based On Statistics

Along with the improvement of the consciousness on security, encryption techniques have been used broadly in the data that is backdoor and is called for Intelligence. The search of data encrypted is so difficult and time-consuming in huge network, especially when the characters of data are unknown. The issues have not been solved though some units have been trying to find a substitution. Known-characters data which is encrypted can be identified efficiently with the use of existing analysis techniques. However, it is difficult to identify data whose characters are unknown. The cipher randomicity checking algorithm is innovatively applied to identify the data encrypted. In this paper, we pay attention to mathematic statistics on testing data and focus on the characters in it. And we gain some performance.The background and the main topic of this thesis are presented in the introduction. Also, it contains the status in the respect of Encryption Identification and some advanced application which has been adopted in data identification.In chapter 2, encryption theory is introduced. The encryption architecture comprised three aspects: encryption based on confusion and diffusion; encryption based on mathematics; encryption based on chaos. We describe the important step of encryption and make a conclusion of the character of cipher-text with different cipher. A support can be provided for the following encryption identification arithmetic.Chapter 3 and Chapter 4 discuss theory and implementation of arithmetic. Encryption recognition algorithms based on mathematical statistics are analyzed about balance, vibratility, periodicity, compressibility, linear dependence, as well as local balance. Next, we apply the means which contain information entropy calculation, run-length statistics, local accumulation as well as global statistics for the data for recognition. Improvements are made on some of these algorithms and such improvements can distinguish between encrypted data and non-encrypted data much effectively. Such idea breaks the traditional way in which we obtain data by characteristics from itself.Chapter 5 presents data encrypted identification methods in the network based on statistics, and experimental results are analyzed. On the purpose of obtaining experimental results curve, adjusting the length of different data and algorithms combinations in experiment are studied in several typical data encrypted. Studies have been completed to make guidance on identification of encrypted data and non-encrypted data.As respect to acquisition and processing of data encrypted, we have different analysis means in the circuit layer, in the signal layer also in information layer. This idea is regard as a supplement to of the current analysis methods in the information layer. The ultimate goal of this method is to provide a solution in effectively recognizing the data encrypted with an unknown type than other means.