FAST LOCALIZATION OF FPGA PROGRAM CODE ELEMENTS SUITABLE FOR STEGANOGRAPHIC STORAGE OF ADDITIONAL DATA

Abstract

This article is devoted to the use of a steganographic approach to storing control data used in the process of hidden monitoring of the security characteristics of FPGA program code. With this approach, the control data forms a single whole with the information object of the program code, and the very fact of the presence of this data in the program code is not obvious. Traditionally, methods based on equivalent transformations of software code elements are used for steganographic embedding of additional data into FPGA program code. However, the volume of data storage provided by such methods is often insufficient for the simultaneous storage of control data for several types of monitoring. Another group of steganographic data storage methods is based on non-equivalent transformations of program code. These methods are used for computing elements implemented in an FPGA environment that perform approximate data processing. The structure of these computing elements includes elementary FPGA units that calculate only insignificant digits of the results of approximate operations. The program code of such units can be changed in a non-equivalent way without causing the device to malfunction. Known methods of this class are based on simulation, which finds all insignificant units, but they are significantly time complexity. This paper proposes an approach that allows finding an incomplete set of insignificant blocks, but with a significant reduction in time complexity. An experimental study of the proposed approach shows that its use allows for an increase in the volume of steganographic storage of control data in the FPGA software code environment. The proposals of this work are applicable to computing elements whose complexity does not allow for the localization of insignificant elementary units using existing methods.