Ubiquitous Knowledge Bases for the Semantic Web of Things

1. INTRODUCTION The Semantic Web of Things (SWoT) is an emerging vision in Information and Communication Technology, joining the Semantic Web and the Internet of Things. The Semantic Web initiative (Berners-Lee, Hendler & Lassila, 2001) aims to allow software agents to share, reuse and combine information available in the World Wide Web. The Internet of Things vision (International Telecommunication Union, 2005) promotes on a global scale the pervasive computing paradigm, which aims to embed intelligence into ordinary objects and physical locations by means of a large number of heterogeneous micro-devices, each conveying a small amount of information. Consequently, the goal of the Semantic Web of Things is to embed semantically rich and easily accessible information into the physical world, by enabling storage and retrieval of annotations from such tiny smart objects. The SWoT vision has significant impact also on human-computer interaction models, with the goal of reducing the amount of user effort and attention required in order to benefit from computing systems. This applies also to computer-computer interaction: traditional Internet-based applications (email, Web browsing, VoIP) adopt a node-centric conversational model, allowing a host to address another one and establish a connection. In contrast, in pervasive computing user agents –running on personal mobile computing devices– should be able to interact simultaneously with many embedded micro-devices and other network nodes, extracting data from objects deployed into the environment. Such data should be automatically processed in a context-aware way, in order to better support the current activity of a user and satisfy her needs in an automated and unobtrusive fashion, with the user even not necessarily aware of specific device interactions. Consequently, the SWoT calls for pervasive knowledge-based systems that achieve high degrees of autonomic capability for information storage, management and discovery, also providing transparent access to information sources in a given area. This implies necessarily an information-centric (a.k.a. data-centric, content-centric) network infrastructure (Jacobson et al., 2009a), where information resources, not hosts, are at the core of protocol design. While research in sensor networks has already shown (Intanagonwiwat, Govindan, & Estrin, 2000) that a content-centric approach is more effective than a node