A Comparison of GIS Architectures for Implementing Indoor Location-based Services

This paper describes our experiences and lessons learnt in building a Geographic Information System (GIS) specifically designed for indoor location-based services. The proposed system was built by mashing-up commodity Open Source software and novel research prototypes realized within our labs. The overall approach relies on intense usage of standard web technologies and REpresentational State Transfer (REST) APIs as a way of enabling easy mashup of off-the-shelf and proprietary components. The key design and implementation aspects of our solution are described in detail, including a discussion on how we represented and augmented the concept of indoor location within our services, as well as on how we integrated them with commodity GIS services originally designed for outdoor scenarios. Further, we show different approaches for implementing client-side applications capable to interact with our indoor location-based services, and we report experimental results of test campaigns for evaluating the impact of those approaches on resource constrained devices.