Data driven security alarm model for embedded applications

The main objective of the work is to propose a security alarm model for embedded crypto application considering the latency value of the process by different computational techniques. An embedded crypto application using Vernam ciphering is considered which may give an unexpected erroneous output in real time environment due to transients and deliberately injected faults. An advanced proactive security alarming technique is proposed for the design and development of such embedded application which will indicate secure and safe operations by prediction. The nature of the attack and its corresponding security measure of the embedded application is modeled using embedded system modeling language (ESML) and implemented in VLSI platforms. The probability of getting a security alarm in run time condition depends on the design parameters like latency, code version and the hardware platform on which it is running. The application is modeled as an event based one and it is monitored based on the design data from the developer. The probability of the proposed security alarm is verified in the design phase when running under different platforms in different versions with redundant processes.