Managing adaptive pervasive computing using knowledge-based service integration and rule-based behavior

D. Lewis, O. Conlan, D. O'Sullivan, R. Wade
2004 IEEE/IFIP Network Operations and Management Symposium (IEEE Cat. No.04CH37507)  
The commonly articulated vision of pervasive computing represents a huge increase in the number of independently developed components that must interoperate and in the level of autonomy they must demonstrate. This motivates a shift to exchanging interoperability knowledge between components at runtime coupled with the ability of components to dynamically adapt themselves to the requirements of users moving between spaces and tasks. The confluence of the service-oriented techniques and
more » ... iques and ontology-based semantics as Semantic Web Services, offers such dynamic adaptivity through knowledge-based service composition. We aim to establish a conceptual architecture for pervasive computing that integrates semantic service composition and policy-based management in providing collective behavior that adapts to the user's changing needs, but which conforms to the goals of those responsible for the resources used by those services. This architecture must remain simple enough to have the potential for ubiquitous deployment over restricted computing resources and be flexible enough to accommodate the rollout of new automated reasoning techniques. The core feature of our architecture is the Adaptive Service Element (ASE), which is characterized by a service description, a model of the state observable by the ASE, a description of the services it makes use of and a rule-based model for describing and restricting its behavior. The lifecycle of an ASE is managed primarily through the bindings made between these models. Service Descriptions are specified in OWL-S language ( where a service is described using a description logic ontology specifying inputs, outputs, preconditions and effects. Compatible with the OWL-S service description, the ASE state view is an ontological model of the objects of which the element is aware, including managed resources, external context and operational state such as counters and timers. An ASE's behavior rules are in the form event-conditions-actions and dictate the behavior of the service in reacting to: the service's invocation, the access control policies of the service provider, the resource management policies of the resource owner and changes to state objects. Meta-rules, typically established by the ASE developer, restrict how available events, conditions and actions can be constructed into behavior rules, thus restricting unwanted rule behavior. ASEs may be generated by automated service composers and thus have entirely rule based
doi:10.1109/noms.2004.1317790 dblp:conf/noms/LewisCOW04 fatcat:44qbnimwpjgjzlcwab6lnzwswq