Temporal context model using extended RDF triple for enhancing system performance

Abstract

Context-awareness is the center of emerging ubiquitous and pervasive computing paradigm. Context-aware computing is a computing paradigm to activate services and change them appropriate to given context. Context model is the foundation of context-aware computing because it defines how the system understands the domain and processes contexts as data. Ontology-based context model is widely accepted in ubiquitous computing area because of its simplicity in format, interoperability, and support for inference. However, most current context modeling techniques do not consider time element as their foundation construct though time is an important context itself, and temporal property is basic information for history, order among context occurrences, further inference and learning context, and so on. There are two methods used to add temporal element in context model; ad hoc approach and temporal ontology approach. Ad hoc approach puts time tags into context model definition in a specific way. It has problems in interoperability and extensibility. Current temporal ontology approach has space and performance problem due to triple reification. To reify a RDF triple statement for temporal information, we need ten additional triple statements; four triples to reify basic triple, and six more triples to represent temporal information. The reification also degrades performance because it results additional join operations over base contexts and temporal element triples. Current temporal ontology methods using RDF triple have cost and semantics problems. The problems are significant so that temporal context management cannot take in dynamic environments. The problems are discussed using mathematical analysis and literature study in introduction and related work. Therefore, we propose a temporal ontology model using quintuple