Mixed Copy Of The Management Of The Grid Environment Topology

Large-scale computers have been utilized to solve various complicated problems since the 1980s. However, in the spheres of science, engineering, and commercial computation, there are still lots of problems beyond the reach of current super-computers. For example, resources of liquor capacity data waiting for treatment are distributedin diferent places, required specific counting and input, output equipments, etc. are not local, and so on. Considering above cases, the Grid Computing Technology is proposed.The Grid is an environment integrating computation and resources, namely, a computing resource pool. The Grid is capable of fully absorbing various computing resources, and then transforming them into a kind of computing ability that is available everywhere, reliable, standard, and meanwhile economical. Besides computers of different variety, the computing resources here also contain Internet communicating capacity, data, instruments, equipments, and even human resources, etc. The problem based on the Grid is the Grid Computing.Currently, the replcia management of existing Grid project use static, or application-driven and user-driven replication rvices, and can not adapt to dynamic character of Grid. This paper presents a new replica management mechanism(hybrid grid replica management topology HGRMT). In HGRMT, it combines both the ring and tree topologies.At the end of the paper, it introduces some simulation experiments about HGRMT. The results show that HGRMT have good scalability, low latency, reliability, adaptability and ease of implementation.