Security Analysis Of Mobile Internet Applications Code Similarity

In recent years, mobile Internet develops rapidly, there are a variety of mobile Internet applications emerging in the market. However, there are some criminals modify the official application illegally and then re-signed it for their own interests. Finally upload the modified applications to third-party market for unsuspecting users to download.This behavior has endangered the interests of the users and stolen their information. At present, the development of mobile Internet applications mainly focused on Apple’s iOS system and Google’s Android system.Different from iOS system, Android system is an open source project.Which the above problems are much more rampant in the Android system.At the same time, there are a large number of third-party Android mobile Internet application stores in China. Compared to the official store, the censor standard is much less strict in the third-party store. According to statistics, most of users download their mobile applications from the third-party stores. The users on Android system face far greater security risks then other mobile operating system platfonn users. Based on the above analysis, the paper will make a code similarity security analysis of mobile internet applications and the research targets are set to Android mobile Internet applications.After research, the current situation of Android mobile Internet applications similarity detections method mainly focuses on the detection of application activities similarity. And there are two deficiencies about this method: one is the feature extraction. The feature model is difficult to accurately describe the characteristics of Android mobile Internet applications and the extraction algorithm is too complex. Another is low efficiency on processing. The amount of Android mobile Internet applications is huge and the distribution characteristics is widely.Complex feature extraction and comparison algorithms, especially feature extraction techniques based on dynamic analysis, has great time costs.Which make the similarity detection of applications lose it practical value.Android mobile Internet application similarity detection requires more accurate feature model to describe the detection of application samples while costing less time during detection. The paper provided a detect method of Android applications similarity based on the abstract syntax tree. The method will decompile the detected application to get the Java source code and get the abstract syntax tree through syntax lexical analysis.It would also caulate the hash value of the tree node according to its syntax characteristic to slove the problems both of complex feature extraction and keep the structure information even after confusing. In comparison pharse, the method will convert the tree node from nonlinear storage model to linear storage mode. The comparison will work between nodes one by one which avoid the mutli travering of the syntax tree and improve the effiency. Thourgh experiment comparison, the method has reached a high accuary.