中文摘要：

针对集中式机制的空间数据服务发现的各种弊端,如单点失效、扩展性差等,提出一种分布式空间数据服务发现机制。基于信息检索模型来实现分布式的空间数据服务发现,将空间数据服务元数据分别表达为索引词汇和地理范围,其中,索引词汇表示服务的非空间属性;地理范围表示服务的空间属性,并提出索引词汇相似性和地理范围相似性来计算查询请求与服务对象间的匹配程度。最后给出了该发现模型在P2P网络的设计与实现,提出一种节点相关模型来将空间数据服务聚簇成不同的地理语义组,查询协议在组间采用有偏漫步试探性查询,在组内则采用泛洪机制查询。实验表明,上述分布式空间数据服务发现模型具备良好的可行性;基于该模型在P2P网络上的实现系统性能较好,是有效克服集中式空间数据服务发现机制各种弊端的途径之一。

英文摘要：

Centralized mechanism for geospatial data services discovery has many drawbacks, such as a single point of failure, poor scalability, so we present a new distributed mechanism for geospatial data services discovery based on information retrieval models. In our models, the metadata of a geospatial data service is described by some index terms and one geospatial bound, here the index terms are non-spatial distributes of the service and the geospatial bound is spatial distribute of the service. And then we provide index term similarity function and geospatial bound similarity function to match query requests against geospatial data services. In our imple- mentation over P2P network, we put forward a peer similarity model to cluster geospatial data services into different geospatial semantic groups. During querying, the query request biased walks between groups and then floods in the target peer＇ s groups. The experiments show that our distributed model for geospatial services discovery has good feasibility and our implementation over P2P network aLso has high performance.