Design optimization of secure message communication for energy-constrained distributed real-time systems

摘要

Modern Distributed Embedded Systems (DESs) tend to be more connected to other peers or external networks. This opens the gate for potential security attacks, although it is beneficial for embedded applications running on DESs. Security protections in DESs require significant time and energy overhead. Longer execution time rises the possibility of deadline violation of real-time applications, which may cause serious consequences in hard real-time systems. Energy efficiency is required since energy budget is usually very scarce under the context of DESs. Therefore, it is of critical importance to study the trade-off between security protection and corresponding timing and energy overheads so that the overall system performance can be maximally improved. In this paper, we approach the design of energy- and security-critical distributed real-time embedded systems from the early mapping and scheduling phases. We focus on providing the best confidentiality protection of internal communication in DESs under timing and energy constraints. The complexity of finding the optimal solution grows exponentially as the problem size increases. Therefore, we propose an efficient, genetic algorithm based heuristic to solve the problem. In the presented algorithm, solutions are evolved gradually so that good solutions can be obtained. Extensive experiments, including a real-life case study, demonstrate the efficiency of the proposed technique.