Design And Implementation Of Distributed Fault - Tolerant System Based On

Related researches at home and abroad for the distributed system have been carrying out for a long time. Distributed system has the characteristics of the platform diversity, people with the introduction of midware platform for the way to solve the heterogeneity and interoperability. Before the introduction of midware, people will be a lot of energy and time to deal with environment, such as cross-platform, across the interoperability problem, to consider the credibility of the system. At a time when the question really is not outstanding, but with the ability of the CPU and Internet broadband technology, the development of wireless network technology, real time and reliability requirements for distributed systems is higher and higher.The distributed system is widely used for its outstanding advantages such as robust, easy scalable, quick response and shared resource. With the development of technology and the expansion of application field, more and more distributed tasks have strict requirements on QoS. In the distributed system, we guarantee the real-time attribute of tasks by the measures supporting fault-tolerance. So when designing new pattern distributed system with high requirements, we should enhance fault-tolerant performance of the system emphatically on the basis of the platform adaptability and the system scalability, thus improving the reliability of the system.This article starts from the view of engineering, to explore the characteristics of distributed fault-tolerant system on object level, the main work is as follows:1. With TAO real-time communication middleware which corresponds the CORBA specification, the object level redundant construction of fault-tolerant system based on TAO, which make the system independent of programming language, operating system and communication protocols. In the mean time, real-time performance has very high guarantee.2. Hierarchical fault tolerant model, facilitates the system from top to bottom, from distributed to a single point, from automatic system fault tolerance strategies to establish a set of comprehensive manual intervention, effectively enhance the reliability of the system.3. For the key object and frequent calls, using non-main-slavery redundancy mode, object and its copy serve at the same time, reducing system load in the prerequisite of ensuring fault tolerance.