Software Execution through OpCode

Abstract—Previous studies on estimating software reliability employed statistical functions for next system failure prediction. These models used parameters based on assumptions regarding the nature of software faults and debugging process. However, none of the existing models, attempted on ensuring reliable runtime system operation. To serve the current demand of autonomous, reliable, service-oriented software, we present a novel approach for runtime reliability estimation of executable software. The approach can help control software execution at runtime by monitoring software state-to-state transition at runtime. The approach involves representing executable software as an automata using opcode extracted from executable code. The extracted opcode is then used to learn stochastic finite state machine (SFSM) representation of executable software which is later employed to trace software state-to-state transition at each runtime instance. An evaluation of our approach on Java-based Chart generator application is also discussed to explain how we can ensure reliable software execution and prevent software failures at runtime with the proposed approach