| With the development of computer science technology and the network application, the information science and technology has entered into many areas of human lives. As the most typical real-time system, the distributed real-time system has been playing an important part in various areas such as the industrial control system, weapon defense control system, the flight control system, the power station control system and the real-time multi-media system, besides; the real-time fault-tolerant technology also obtained thorough research.On the basis of redundant resources, the real-time fault-tolerant technology uses technologies such as fault-tolerant detection, diagnosis, isolation and fault-recovery to guarantee the real-time system to provide correct services while faults break up via managing the software and hardware structure of the system reasonably, which was proposed to solve the fault-tolerant problems of real-time system.This paper designed the distributed real-time fault-tolerant system based on fully analysis and researches of present resolution of distributed real-time system and its fault-tolerant technology. The system was designed under Linux, all the nodes will be divided into several groups, and the nodes in the group apply the master-salve structure. Nodes could enter or exit a group by sending the corresponding messages and every node in this group will run the system respectively and provide monitoring and protection to each other.Compared with other Fault detector we realize the heartbeat protocol under the kernel level and schedule the heartbeat packets by an hard real-time protocol—TTEP, which has shorter detect period and meets the need of real-time better. Design a distributed real-time task scheduling module as the core scheduling algorithm of this real-time fault-tolerant system, which could schedule the takes and make the corresponding adjustment according to the running statement of the system on the basis of considering various performance indicators. |
PDF Full Text Request