Abstract
Real-world information systems (such as those dealing with RFID, sensor networks, etc.) could really benefit from a real integration with the Web. Current integration projects often focus on using the Web as a transport layer (e.g. WS-* services) rather than as an applications protocol. A RESTful approach for these systems would make them accessible to a broad community of Web developers using immensely popular Web languages (e.g. HTML, URI, JavaScript, etc.) and leveraging well-proven Web mechanisms (e.g. browsing, searching, bookmarking, caching, and linking). To emphasize this fact and better grasp the differences between the WS-* and RESTful integrations schemes we will propose a concrete case study: the EPC Information Service.
The Electronic Product Code (EPC) Network is a set of standards established by industrial key players towards a uniform platform for tracking and discovering RFID (Radio Frequency Identification) tagged objects and goods. This network offers, amongst other components, a standardized server-side EPCIS (EPC Information Service) which is in charge of managing and offering access to track and trace RFID events. Implementations of the EPCIS (such as the open-source Fosstrak platform) offer a standard query and capture API through WS-* Web services.
In this talk we will show how we translated this standard WS-* interface into a RESTful system leveraging both HTTP as an application protocol and the ATOM Web publishing protocol. We will then discuss the advantages and drawbacks of each approach, focusing both on quantitative facts (e.g. performance of the system) as well as on qualitative facts (e.g. ease of programming).
We will then show how the RESTful version of the EPCIS enabled us to create a dashboard in the form of a simple mashup application. Using this dashboard the tagged objects can be traced on Google Maps and Earth and looked for on Wikipedia. The inventory can be visualized in real-time using Google's Visualization API.
Finally, we will explain how the RESTful EPCIS was developed as a JAX-RS (JSR-311) application using Jersey and running on Tomcat or Glassfish V3, and demonstrate its integration with the Fosstrak opensource platform.