Semantics-based interoperability for pervasive computing systems

The openness and dynamics of the pervasive computing environment call for dynamic integration of autonomous, heterogeneous networked components. This position paper outlines an approach towards ad hoc interoperability in pervasive environments. 1 Pervasive Environments and Promising Paradigms The paradigm of pervasive computing [1] has opened new perspectives to the enactment of human activities related to accessing information and computation. Pervasive computing enables user-centric retrieval and consumption of information, compared to the conventional computer-centric approach. Systemically, this is realized as a synergistic combination of intelligent human-machine interfaces and ubiquitous computing & networking. The ubiquitous property implies a useful, pleasant and unobtrusive presence of the system everywhere around us. Pervasive computing systems are characterized by a number of features, such as: (i) the highly dynamic character of the computing & networking environment due to the intense use of the wireless medium and the mobility of the users; (ii) the resource constraints of mobile devices, in terms of CPU, memory, storage, display capabilities, battery power and bandwidth; and (iii) the high heterogeneity of integrated technologies in terms of networks, devices and software infrastructures. To cope with such dynamics and diversity, pervasive computing systems should have the capacity to be deployed and execute in an open, ad hoc way, integrating the available hardware and software resources at the specific time and place. This is facilitated by rendering such resources autonomous, networked components that may be incorporated in a larger system. Then, supporting interoperability between heterogeneous system components becomes a key issue. Many of the network interoperability aspects can be addressed by reliance on the ubiquitous Internet’s network and transport protocols. However, above the network level, the interoperability problem remains. Considering the large number of players and technologies involved in realizing pervasive computing systems, solutions to interoperability based on reaching agreements and enforcing compliance with interoperability standards cannot scale. Instead, networked system components shall adapt at runtime their functional and non-functional behavior in order to be composed and interoperate with other components. Moreover, supporting composition and interoperation requires the definition of behavioral conformance relations to reason on the correctness of dynamically composed systems with respect to both functional and non-functional properties. Pervasive computing systems have incorporated various software technologies and development models coming from the mature domain of distributed computing systems and further specialized for mobile computing systems, e.g., middleware for distributed objects. A recent such computing paradigm which seems particularly appropriate for pervasive computing systems is Service-Oriented Architectures (SOA) [2]. Therein, networked devices and their hosted applications are abstracted as services delivered and consumed on demand. The benefit of this approach lies in the looser coupling of the components making up a system, hence the increased ability to making systems evolve as, e.g., requirements change or the computing & networking environment changes. The SOA approach, as, e.g., enabled by the Web Services Architecture, appears to be a convenient architectural style enabling integration of pervasive services that may dynamically be retrieved and composed, thanks to service discovery and coordination protocols. However, the SOA paradigm alone cannot meet the interoperability requirements of pervasive computing systems. Drawbacks include: (i) enforcement of a specific middleware platform to ensure integration at the communication level; and (ii) interaction between services based on their syntactic description, for which common understanding is hardly achievable in open environments. Then, a promising approach towards addressing the interoperability issue relies on semantic modeling of information and functionality, that is, enriching them with machine-interpretable semantics. This concept originally emerged as the vehicle towards the Semantic Web [3]. Semantic modeling is based on the use of ontologies and ontology languages that support formal description and reasoning on ontologies; the Ontology Web Language (OWL) is a recent recommendation by W3C. A natural evolution to this has been the combination of the Semantic Web and Web Services into Semantic Web Services [4]. This effort aims at the semantic specification of Web Services towards automating Web services discovery, invocation, composition and execution monitoring. The Semantic Web and Semantic Web Services paradigms address application-level interoperability in terms of information and functionality [5, 6]. However, interoperability requirements of pervasive computing systems are wider, concerning functional and non-functional interoperability that spans both the middleware and application levels; conformance relations enabling reasoning on interoperability are further required. Work in the field of Software Architectures [8] has provided the basis for reasoning on the correctness of dynamically composed systems with respect to both functional and non-functional properties, at middleware and application level. One such effort, described in [7], elaborates base modeling of mo1 http://www.w3.org/TR/ws-arch/ 2 http://www.w3.org/2001/sw/ 3 http://www.w3.org/TR/owl-semantics/ bile software components that integrates key features of the mobile environment, to support correctness of dynamic composition. It is clear from the above discussion that interoperability in pervasive computing systems is a major and still open issue. This position paper outlines an approach towards ad hoc interoperability in open pervasive computing environments, building upon research background coming from the domains of Service Oriented Architectures, Semantic Web Services and Software Architectures. 2 Approach to Ad Hoc Interoperability As introduced in the previous section, service orientation is an architectural paradigm that provides many advantageous features for ad hoc system composition; it is, however, restrictive in imposing a specific middleware technology. To remove this restriction, we introduce an abstract, middleware technology-independent reference system architecture, which, nevertheless, retains the desirable characteristics of service-orientation. heterogeneous middleware layer heterogeneous platform layer service communication heterogeneous application layer syntactic service description functional specification context specification QoS specification