OGC Web Map Service Implementation Challenges for Mobile Computers Clodoveu A. Davis Jr. Departamento de Ciência da Computação Universidade Federal de Minas Gerais Belo Horizonte, Brazil clodoveu@dcc.ufmg.br Yuri Jorge Kimo, Fátima L. P. Duarte-Figueiredo Instituto de Informática Pontifícia Universidade Católica de Minas Gerais Belo Horizonte, Brazil yurikimo@gmail.com, fatimafig@pucminas.br Abstract—One of the greatest challenges for computing today involves providing access to large volumes of distributed multi- media data. Mapping and personal navigation are among the most interesting applications involving this challenge, since they require access to static (such as street maps) and dynamic (such as traffic conditions) geospatial data. Such applications are cur- rently the focus of much attention, considering that hardware is improving and wireless networks are becoming ubiquitous. Access to online maps is mainly done through proprietary and thematically limited geographic information services, such as Google Maps. A more interesting alternative for mobile applica- tions is to use sources from spatial data infrastructures based on the Open Geospatial Consortium’s standardized Web services, such as the Web Map Service (WMS). WMS, however, imposes a high communications overhead and power consumption to the mobile device, and has a limited scalability potential. This paper presents a proposal for the development of an SDI environment in which some of these challenges can be adequately met, from both a geoinformatics and a mobile computing perspective. Keywords- geospatial web services; geographic information systems; mobile GIS; Web Map Service; spatial data infrastructures I. INTRODUCTION The convergence of mobile computing with Web-based re- positories of geographic data and geotechnologies, such as the Global Positioning System (GPS), favors new mobile geo- graphic applications. Lower hardware prices and the integra- tion of GPS into mobile phones and PDAs have provided many people with precise and costless personal location information most of the the time. Using ubiquitous wireless networks, such as WiFi, GPRS, and 3G, users now demand better location- based services and applications. Most geographic applications available today for mobile phones and PDAs are based on static geographic datasets, ei- ther downloaded to the device or available through regular Web access to mapping sites. Even though such datasets are important and functional, users want to select among other available datasets, through which they can fulfill specific needs, gain access to more specialized local data, and combine various information sources for better decision-making. In the commercial mobile computing arena, geographic in- formation services are still proprietary and limited. Companies such as Google, Yahoo!, Garmin, and Nokia have developed solutions that offer digital map browsing and provide simple mobile applications such as routing and locating nearby points of interest. These applications need constantly updated data, reflecting today’s intensive urban dynamics. Such updating is not easy, much less for global companies such as the ones mentioned. As a result, online maps and applications always fall behind reality in many aspects, thus leading to errors and loss of user confidence [9]. The integration of new data in a Web map constitutes a problem, although it is possible to create mashups by plotting thematic data on top of a generic Web map backdrop using JavaScript-based APIs. Information sources for that purpose, however, remain scattered all over the Web, and it is not sim- ple to discover and integrate many of them, in a way that would allow users to select and activate/deactivate new data as layers in a desktop GIS. A more interesting solution for disco- vering and accessing various geographic information sources through the Web involves the use of service-oriented architec- tures (SOA). In SOA, data providers publish metadata on available datasets using a catalog service. Clients search this catalog, identifying services of interest (an activity known as discovery). The client can then connect (bind) to the selected service(s), get from them a list of available data classes (or layers), and select which ones are required [8, 26]. The SOA approach to data sharing has been conceived as an interoperability solution, initially by the Open Geospatial Consortium (OGC) [21], then redeveloped by the W3C [27]. This approach has countless advantages. Clients do not need to know in advance which data is available at which server, since that information is kept in the catalog. Servers can go offline, and replicas can take their place by simply changing the catalog. If there are no replicas, the user can select a simi- lar service as a data source. Clients do not need to install spe- cialized software, a simple Web browser often suffices. The precise format in which data is coded and stored becomes irre- levant; data transfer between client and server uses the Geo- graphic Markup Language (GML), an OGC standard based on XML, thus technologically neutral. Furthermore, database managers, application services, and viewing applications can be developed independently from one another, relying only on the standardized interface between them. Services are specialized by the OGC according to the type of data they provide. The most important ones are the Web Map Service (WMS), the Web Feature Service (WFS) and the Web Coverage Service (WCS). WMS produces complete maps, rendered at the server and transmitted as images to the