Minimizing the side effect of context inconsistency resolution for ubiquitous computing

Abstract

Applications in ubiquitous computing adapt their behavior based on contexts. The adaptation can be faulty if the contexts are subject to inconsistency. Various techniques have been proposed to identify key contexts from inconsistencies. By removing these contexts, an application is expected to run with inconsistencies resolved. However, existing practice largely overlooks an application's internal requirements on using these contexts for adaptation. It may lead to unexpected side effect from inconsistency resolution. This paper studies a novel way of resolving context inconsistency with the aim of minimizing such side effect for an application. We model and analyze the side effect for rule-based ubiquitous applications, and experimentally measure and compare it for various inconsistency resolution strategies. We confirm the significance of such side effect if not controlled, and present an efficient framework to minimize it during context inconsistency resolution. © 2012 ICST Institute for Computer Science, Social Informatics and Telecommunications Engineering.