网格环境中分层次哈希表资源发现机制  

Level hash table resource discovery mechanism in the grid environment

作　　者：王玥[1] 刘卫东[1] 宋佳兴[1] 

机构地区：[1]清华大学计算机科学与技术系,北京100084

出　　处：《清华大学学报（自然科学版）》2006年第4期543-546,共4页Journal of Tsinghua University(Science and Technology)

基　　金：国家自然科学基金资助项目(90104002;90412012);国家"九七三"重大基础研究项目(G1999032707)

摘　　要：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。In the grid environment, resources are heterogeneous dynamic, and widely distributed, so the resource discovery becomes a challenging task. A resource discovery mechanism called the Level Hash Table mechanism was developed using a tree-based network architecture to faciliate resource usage. In general, resources are hashed into keys with each hash table node storing keys that can be mapped to the registered resources. A simulator was built to model the mechanism with up to 700 nodes. The results show that the resource discovery process can finish in O（logqp） hops and that the number of child nodes significantly influences the performance. The results also show that the level Hash table mechanism performs better than the traditional flooding mechanism.

关 键 词：网格 资源发现 分层次哈希表 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]