类型感知的多属性资源查找机制

Type Aware Multi-attribute Resource Discovery Mechanism

针对网格环境的分布性、动态性和资源异构性等特点,在基于超级节点的层次化结构中,提出了一种类型感知的多属性资源查找机制。与传统的节点分组方式不同,提出了资源分组的思想,依据资源类型以及资源的相似程度对进资源进行分组,保证同一分组内资源类型相同,属性相似,使得同组内资源的"纯度"较高,为提高资源查找效率创造条件。同时,在资源的管理方面,为每一个资源分组分配管理节点,负责资源注册、资源匹配等方面的工作。仿真实验表明,该类型感知的多属性资源查找机制至少能够激昂平均跳数、响应时间。消息数分别降低15%,15%,25%。同时,并且随着系统规模的增大,性能仍得以保持。同时,该机制能够在一定程度上提高查找的精确性。

To get over the resource instinctive characteristic of decentralized,dynamic and heterogeneous,a type aware multi-attribute resource discovery mechanism was proposed.Unlike the traditional node grouping,the mechanism deploys the idea of resource grouping according to the resource type and attributes similarity,which ensures that resources within a group share the same resource type and they are similar to each other.This partition approach makes the resource group ＇purer＇,which will lead to accurately searching.Meanwhile,there is a manage node for each resource group who is responsible for resource register,matching and maintenance.Simulation shows that the type aware multi-attribute resource discovery mechanism can decreases average hop counts,response time and messages brought by resource forwarding by at least 15%,15%,25% accordingly,and the performance can be preserved when the system scales up.Simulation also shows this mechanism can increase the accuracy of searching in a certain degree.