Secure cloud model for intellectual privacy protection of arithmetic expressions in source codes using data obfuscation techniques

Abstract

The source code of any software is confidential and its intellectual property has to be protected from the malicious intent of the attackers. The present work offers the architecture of a cloud model for secure intellectual property protection of arithmetic expressions in a source code. The data obfuscation techniques are meant to be applied on different types of arithmetic operators in an arithmetic expression, “The required algorithms proposed in this paper are named as i) ‘Adaptive Number Generation’ obfuscation to be applied on all operators ii) Obfuscation by ‘Carry Padding’ to be applied on ‘+’ operator iii) Three Obfuscations by ‘Mixed Boolean Arithmetic’ to be applied on ‘+’, ‘-’ and ‘*’ operators and iv) ‘A-RSA+′ obfuscation to be applied on ‘+’ operator.” Experimentations are conducted to validate their functionality and performance metrics measurements in terms of Cyclomatic complexity, Time complexity, Space complexity and Potency analysis are carried out on a set of benchmark source codes and it has been found that there is only a very slight increase of 6.57% number of lines of code in the obfuscated codes compared to the corresponding original, the average increase in the obfuscated code's execution time is 29% and the estimated average potency given by the proposed obfuscation techniques is 1.06% which is also greater than one. Also, a mathematical model is developed to estimate the Mean Obfuscated Source Code Cloud Delay (MOSCD), which is the time duration in which the obfuscated source code is available physically at the cloud server to face any possible Man-At-The-End (MATE) attacks. An attacker may build a control flow graph (CFG) out of the source code to perform attacks based on static and dynamic analysis-based reverse engineering. The time required to generate a CFG is also estimated and in comparison, with the MOSCD, it is observed that the minimum CFG build time is almost three multiples of the maximum MOSCD value we received in our experimentations it is proved that our model is strong to thwart away any possible MATE attacks.