Assessing data dissemination strategies within Triple Spaces on the Web of Things

The upcoming of the Web of Things initiative has improved the integration of Internet-connected devices through the standard HTTP protocol and other web techniques. Unfortunately, it usually defines the data shared by these devices in a syntactic level, showing a lack of expressiveness. During the last decade, the Semantic Web (SW) has aimed to solve these problems by adding logic to the Web to make it machine-understandable and by therefore enhancing the interoperability of the applications using it. The SW is used in the Triple Space Computing paradigm, which proposes a blackboard model where semantically described knowledge is shared between different devices in a completely RESTful, and consequently WoT compliant, manner. This paradigm's shared blackboard can be implemented using many strategies, from centralized to completely distributed. In this work, we compare and analyze the behaviors of these two extreme cases in several simulations which try to represent common IoT scenarios. Finally, we propose an improvement of the completely distributed strategy by enabling the gossiping between devices.