The Research Of Web Services Discovery Based On Semantic

With the widely use of Web Services technology, the number of web services become bigger and bigger, How to find web services quickly and accurately in a large number of services set is a growing question. As an important and crucial part in the web services architecture, web services discovery technology search and create a candidate services set in the register centre based on users'needs. It selects the candidate services set under certain rules, and gets the most suitable service.Traditional web services discovery method which based on keywords matching has deficiency, and low in recall and precision. Researchers introduce semantic-based web services discovery mechanism. However, the semantic description model of existing services only focus on describing services function, lack of a clear description of QoS, which leads to that services we find have no guarantee in QoS. In view of the above challenges, this paper based on studying the existing web services models, introduce QoS into the service discovery model, designed a QoS-based semantic web services discovery system model SWSDM-QoS.SWSDM-QoS model applies the strategy of multi-stage matching to semantic web service discovery, describes the status and function of every stage during the course of service matching, and gives the idiographic algorithm of matching in every stage. SWSDM-QoS model applies augment UDDI registration system, and improves the grades matching algorithm that augment UDDI registration system adopts, makes up the disadvantage that the matching result is rough. Besides that, SWSDM-QoS model applies a QoS-based services selection algorithm in which user can set the QoS weight according to their needs that to make Web services more suitable for user's needs.This paper implements the core of SWSDM-QoS model, and designs a group of data to test model's algorithm performance. The experimental results show that the discovery mechanism used by SWSDM-QoS has a little ascended in precision and greatly improves recall compared with UDDI-based keyword matching.