Context and policy based fault-tolerant scheme in mobile ubiquitous computing environment

Abstract

In ubiquitous computing system, the increasing mobile and dynamic of software and hardware resources and frequentative interaction among function components make fault-tolerant design very challenging. In this paper, we propose a context and policy based self-adaptive fault-tolerant mechanism for a mobile ubiquitous computing environment such as a mobile ad hoc network. In our approach, the fault-tolerant mechanism is dynamically built according to various types of detected faults based on continuous monitoring and analysis of the component states. We put forward the architecture of fault-tolerant system and the context-based and policy-based fault-tolerant scheme, which adopts ontology-based context modeling method and the Event-Condition-Action execution rules. The mechanism has been designed and implemented as self-adaptive fault-tolerant middleware, shortly called SAFTM, on a preliminary prototype for a dynamic ubiquitous computing environment such as mobile ad hoc network. We have performed the experiments to evaluate the efficiency of the fault-tolerant mechanism. The results of the experiments show that the performance of the self-adaptive fault tolerant mechanism is realistic.