基于复杂云资源查询的HChord索引机制

HChord index mechanism based on query of complex cloud resources

为了研究复杂云资源在巨大资源池中快速定位和查找问题,结合分布式对等技术资源定位的优势,提出分层的HChord云对等模型,通过提取云资源多维属性特征向量构建资源查询索引和相似资源聚类,建立全局与局部索引及缓存机制,并独到地提出依赖备份超级节点数来控制索引缓存率的方法。仿真实验表明,HChord模型比HTC-Chord模型在资源定位时需要更短的平均路径长度;验证HChord模型下不同索引缓存率对资源定位路径不同的有利影响。结果表明,分层HChord模型下构建的资源聚类和索引缓存机制,以牺牲少数节点维护开销能使资源在非常有效的路径范围内被定位。

In order to research the problem how to locate and find the complex cloud resources in a huge resource pool quickly,considering distributed peer-to-peer technology’s advantages locating resources rapidly,this paper established a layered HChord equivalent model of cloud,by extracting cloud resources multidimensional attributes vector,it built resource’s searching index and clustered the similar cloud resources to build the global indexes and local indexes and caching-mechanism. At the same time,it put forward the special method controlling the index buffer rate relying on candidate super nodes’ number.The simulation experiments show that HChord model needed the shorter average path length than HTC-Chord when locating resources,and it also verified the HChord’s index-caching rate brought beneficial effects to resource location path. Results show that the HChord model building the index of resource clustering and caching mechanism,can make resources be located within a very effective path by sacrificing a few nodes’ routing overhead.