| High reliable supports are greatly in need among the key applications of complex real-time environment. When the system has the abilities of redundancy and fault-tolerance technologies, it can run quickly, exactly,reliably and continually. How to acquire the abilities in the systems is the major problem to be solved.It is obvious that the establishments and managements of fault-tolerance structure in real-time system software makes the systems more complex. Usually, some special fault-tolerance management software must be worked out according to the redundancy and fault-tolerance control methods. But it will surely increase loads of work, lower the reliabilities of systems and support simplex mode only.In order to solve this problem, a novel fault-tolerance structure was designed, which is called redundancy and fault-tolerance libraries technology. When this technology is used , the four widely used fault-tolerance control mode, including triple modular fault-tolerance, triple modular redundancy (TMR) and so, are supported by the operating system kernel, which also provides the API interface for different modular management and multi-tasks supporting. It can not only keep the system layer streamlined and high efficient, but also reduce the designers' programming workloads, and guarantee the quality.Firstly, survey of the state of the art was given. After introducing some concept about redundancy, fault-tolerance and reliabilities evaluating technologies, the key points of redundancy and fault-tolerance system design is summarized. Then the modular and architecture of redundancy and fault-tolerance system is brought forward, and the key technologies among it is also presented, including: self diagnosis (cpu, I/O,memory),heartbeat diagnosis, reliably communication technology, triple modular voting algorithm, system control restructuring and so on. Finally, this paper illuminates and summarizes the state of affairs of our system, along with the testing results. |
PDF Full Text Request