Arquitetura DyCoRS: uma abordagem para ciência de contexto explorando composição dinâmica de estratégias para raciocínio

Abstract

Due to its high scalability, ubiquitous computing uses management mechanisms that have the most autonomous operation possible, requiring approaches that enable context awareness. Context awareness can be structured in four steps: acquisition; modeling; processing; and dissemination. Among these steps, we highlight the context processing, which, among other features, it is responsible for performing contextual reasoning inferring new knowledge and improving understanding of acquired contexts. Several strategies have been used to implement context reasoning, and a strategy that can be applied satisfactorily in different application scenarios has not yet been found. Because of this, hybrid approaches for context reasoning are gaining prominence. These approaches allow the combination of two or more strategies, being able to take advantage of the positive points and minimize the negative aspects of each strategy. Although these approaches are already being used, we still note a lack of flexibility in making available the resources offered. Thus, this work aims to design an architecture called DyCoRS that is dynamic in providing the necessary resources to obtain context awareness. Among the features offered, we highlight the hybrid approach for context reasoning, which allows the dynamic composition of different strategies. Besides, the architecture also provides: (i) different storage models; (ii) a method for correlating persisted information in the various models; and (iii) a process that allows the reconfiguration of the resources provided by the architecture at runtime. We evaluate the proposed architecture through usage scenarios, exploring in each one of them, different functionalities provided by DyCoRS. The results show that it is possible to offer greater flexibility for context-aware applications to choose how they want to perform the reasoning of contextual information. Thus, the results stimulate further research.