A Spaces-Based Platform Enabling Responsive Environments

Space Integration Services is a software communication platform that enables the seamless integration of sensors, actuators, and application-logic components through a multispace model and a spaces-based publish/subscribe mechanism. The underlying model is based on finite spaces only, thus the application-logic components that need to reason on spaces like the geodetic or the Cartesian ones, are obliged to use a discretized version of those spaces and to maintain the correspondence between “real” locations and their discretized representation if they are interested in exploiting the services offered by the platform. To compensate for these limitations, the platform has been extended in order to add support for spaces with an infinite number of locations, (e.g., spaces described by continuous coordinates, such as geodetic or Cartesian spaces). Such extension is essential for the use of SIS also as enabling platform for outdoor pervasive computing systems thanks also to the wide spread of outdoor localization systems. This paper presents the new conceptual model that results from a generalization of the existing one valid for finite spaces only. To encompass both finite and infinite spaces, the new model moves the primitive concept of elementary position from location to spatial context (as a set of locations). Several kind of spatial contexts have been introduced to offer a (quite) complete set of localization typologies, for example, from a single position to a set of positions described by a function. The model has been turned in a prototypal implementation realized by means of an additional layer on top of the finite-space version. Such an implementation has been experimented in a real case scenario dealing with a parcel distribution company. Finally, the performance of the prototype is then compared to the one of the finite-space version in a series of experimental tests.