Abstract. Data service is a specialization of Web service, and end-users can synthesize cross-organizational data by composing data services. As composite schemes overlap each other, some primitive data services could be called repeatedly
by composite data services, so that the response delay and server load are aggravated. In this paper, an optimized strategy for data service response with template-based caching and compression is proposed. Firstly, the strategy uses the message template to extract the application-relevant values from SOAP messages. Secondly, the strategy holds the objects from application-relevant values rather than XML representations to decrease the overhead of SOAP message parsing. Thirdly, the strategy uses the XMill compression algorithm to decrease the volume of data transmitted. Extensive experiments based on Spring-WS-Test benchmark demonstrate the strategy is an effective approach to reduce response latency and server load compared to non-caching tehcnniques.
Keywords: data integration, message template, data caching, data compression, SOAP, data service.
