ProActive-based Parallel Program Fault-tolerant Task-scheduling

Along with the development of distributed computing application, the resource becomes fruitful. The unstable of computers brings down the reliability of distributed computing system. Moreover, malicious hosts and the third entities attacking affect the security of distributed computing system. Correctness and security of computing, which the bottleneck of the development is of distributed computing, has become one of the hot topics of large-scale distributed computing systems.ProActive is a middleware Java library for parallel, distributed, and concurrent computing, also featuring mobility and security in a uniform framework. With a reduced set of simple primitives, ProActive provides a comprehensive API allowing simplifying the programming of applications that are distributed on Local Area Network (LAN), on cluster of workstations, on P2P desktop Grids, or on Internet Grids. In order to guaranteed fault-tolerance and security of the distributed system, we proposed a new fault-tolerant scheduling model which introduction of the trust mechanism, and improved the former distributed computing system called WPHPC. The experiments illustrated that the new fault-tolerant scheduling model can improve the reliability of the system and will not occupy too much computing resource. Main works of this paper as follows:1. Proposed a ProActive-based fault tolerance model, explained its structure and work-theory, proposed and analyzed three kinds of fault tolerance scheduling algorithm. The experiments illustrated that the fault tolerance model and corresponding algorithms make good use of CPU of the computers, ensure the correctness of the execution result.2. Proposed a new fault-tolerant scheduling model which introduction of the trust mechanism. The experiments illustrated that the new model prevented destroy from malicious hosts and the third entities, bought up the utilization of computing node which is of high credibility.3. Based on the model above, improved the former distributed computing system and accomplished a fault-tolerant parallel computing system.