RAFD: Resource-aware fault diagnosis system for home environment with smart devices

With recent advancement in technologies used at home, smart home environment allows various resources such as device, network, or content to be connected to one another. Their configurations are changed by dynamic bindings at any time. In this smart home environment, a minor problem in a resource can trigger serious failures in home network services by causing multiple faults to the related resources simultaneously. To solve this problem, it is essential to analyze the dependency between resources and also to diagnose home network faults autonomously. This paper proposes the effective fault diagnosis system based on resource relation map which is dynamically constructed by information convergence model of heterogeneous home resources. The proposed system provides the tracing method for finding the root cause of a fault using the resource relation map. The resource relation map represents the snapshot of home situations at the given time. The proposed fault diagnosis method allows building cost effective remote maintenance system with high availability and manageability by tracing the fault cause along the dependency between resources using graph-style resource relation map as if humans trace the cause of problem. In addition, it can contribute to realize an autonomic fault management system for smart home. In this paper, the prototype of the proposed system is implemented and evaluated for performance in accuracy and latency of fault diagnosis in a real environment. The experimental results show that the proposed system, especially with the suggested back tracing diagnosis system, yields remarkable performance for home network fault diagnosis.