Accessing Distributed WFS Data Through A RDF Query Interface

GIScience 2016 Short Paper Proceedings Accessing Distributed WFS Data Through A RDF Query Interface Tian Zhao 1 , Chuanrong Zhang 2 , and Weidong Li 2 University of Wisconsin – Milwaukee University of Connecticut Abstract. Geospatial data stored in databases and other formats can be accessed through Web Feature Service (WFS). However, it is not convenient to access data in multiple WFS servers since WFS protocol is geared towards single server. In this paper, we propose an algorithm to query and synthesize distributed WFS data through a RDF query interface, where users can specify data requests to multiple WFS servers using a single RDF query. The algorithm translates each RDF query written in SPARQL-like syntax to multiple WFS get-feature requests, and then convert the WFS results to answers to the original query. A lightweight Web-based prototype is implemented based on this approach. Introduction In this paper, we propose the design of a RDF query interface for distributed WFS data, which accepts queries in SPARQL-like syntax to provide more flexi- bility and usability than direct WFS queries. As a motivational example, imagine a scenario where a developer needs to implement a program to display the flooded streets near the high schools of a city. The programmers can define an interface backed by some predefined WFS queries to fetch data from two servers. For each user request, the interface will request data from the servers and then integrate the results. If the servers use di↵erent data definitions, a translation step is needed to reconcile and integrate the data. If the developer needs to support another query such as finding the bridges of major highways, the developer has to perform the above steps again. While the implementation of two queries shares many similarities, it is not apparent how to re-factor the duplicated code, which include the communication with WFS servers and the transformation and integration of the responses. The difficulty is not with building shared library but with composing specific WFS requests and interpreting and integrating the corresponding results, which may be di↵erent from each query. In addition, while queries may involve the same intermediate data, it is not straightforward to implement a caching strategy to improve performance. This is especially critical during emergency when peak data requests can overwhelm data servers. Our design of RDF query interface aims to improve the productivity for rapid prototyping of WFS query applications. The RDF interface automatically