Research On Task Scheduling Mechanism Of Safety-critical Energy-aware Real-time System

With the rapid development of terminal technology,networked embedded real-time systems have been widely used,and have brought great convenience to human production and life.At the same time,due to the close connection between the embedded real-time system and the internet,the security of the embedded real-time system is threatened.In order to solve the security problem of the networked embedded real-time system,introducing security services are the frequently used methods,such as using encryption algorithms to encrypt some sensitive data of the system.However,embedded real-time systems are mostly battery-powered and have limited energy,encrypting data using encryption algorithms usually consumes extra energy,which may cause the lack of energy for the system to perform other important tasks.Therefore,to ensure the largest security quality of service(QoSS)of the embedded real-time system under limited energy constraints is worth to study.To solve the above problem,the energy limitation and real-time of the systems are considered,and a security-critical and energy-aware task scheduling scheme for real-time system is designed with the objective of maximizing the QoSS of the system,which determines the task-to-core allocation,the operating voltage and frequency of each task,as well as the security services.The main work of this paper is as follows:1)The security problem of the system is constructed as a Mixed Integer Nonlinear Programming(MINLP)problem with the constraints of the energy and real-time.Then,this MINLP problem was transformed into a simple Mixed Integer Linear Programming(MILP)problem with no performance loss.2)An two-stage heuristic algorithm called TTS is proposed to address the problem of high complexity when solving MILP problems.The first stage determines the task allocation and the operating voltage and frequency of the task to minimize the system active energy for task execution.The second phase determines the security services of each task to maximize the QoSS of the system.Experiments show that the TTS algorithm can find sub-optimal solutions in a shorter time than MILP solver and has higher QoSS of the system and higher energy efficiency than other typical algorithms.3)To find a optimization of the MILP problem than TTS algorithm,an improved Benders decomposition method called MPQT is proposed.MILP problem is divided into two small problems with fewer variables and constraints by MPQT: the master problem and the salve problem.To reduce the time complexity,the master problem is relaxed,and a two-level gradient method is proposed to solve the salve problem.By iterating the master and slave problems,the optimal solution then can be found.Comparative experiments show that the MPQT algorithm can improve the QoSS of the system compared to the TTS algorithm.