网格环境中,资源是异质的、动态的,广泛分布于整个网络。基于树型网络结构提出了一种分层次哈希表资源发现机制。资源在此机制中被哈希成为键值,每个哈希表节点存储一系列键值,对应注册在该节点或索引在该节点上的资源。因此,此机制的资源发现过程是有目的的寻找过程。对此机制进行了模拟,系统最大节点数达到700。所得结论:资源发现过程可以保证在O(logqp)步内完成;适当调整树型结构的子节点个数有助于提高性能;该机制性能优于传统洪泛机制。 Grid environment, the resources are heterogeneous, dynamic, widely distributed throughout the network. Based on the tree network structure, a hierarchical Hash table resource discovery mechanism is proposed. Resources are hashed in this mechanism as a key, and each hash table node stores a series of key values ​​corresponding to the resource registered on the node or indexed on the node. Therefore, the resource discovery process of this mechanism is a purposeful search process. This mechanism was simulated, the maximum number of nodes reached 700. The conclusion is that the resource discovery process can be guaranteed to be completed within O (logqp) steps. Proper adjustment of the number of child nodes in the tree structure can improve the performance. This mechanism outperforms the traditional flooding mechanism.