Toward a Configurable Thing Composition Language for the SIoT

With the advent of the Social Internet-of-Things (SIoT), which builds upon the success of IoT, different IoT devices engage in relationships to achieve common goals calling for a composition language that would dictate who will do what, when, and where. Since the number of service interactions may become overwhelming the composition plan may have many alternatives. For that reason, variability is an important concern for IoT and SIoT. A configurable composition model could provide a consolidated view of multiple variants of composition plans. It promotes the reuse of proven practices by providing analysts with a generic modeling approach from which to derive individual composition plan models. Unfortunately, the scope of existing notations for configurable composition modeling is restricted or even non-existent in the IoT domain. In this paper, we propose a variability awareness IoT service Composition Framework. The new Framework handles both configurable and classical IoT service composition. We represent the new primitives, the configurable composition language, and algorithms of customization. Finally we demonstrate the features of the framework through a proof of concept application.