作者: Nithya Ramanathan , Kevin Chang , Rahul Kapur , Lewis Girod , Eddie Kohler
关键词:
摘要: Being embedded in the physical world, sensor networks present a wide range of bugs and misbehavior qualitatively different from those most distributed systems. Unfortunately, due to resource constraints, programmers must investigate these with only limited visibility into application. This paper presents design evaluation Sympathy, tool for detecting debugging failures networks. Sympathy has selected metrics that enable efficient failure detection, includes an algorithm root-causes localizes their sources order reduce overall notifications point user small number probable causes. We describe evaluate its performance through fault injection by active application, ESS, simulation deployment. show broad class data gathering applications, it is possible detect diagnose collecting analyzing minimal set at centralized sink. have found there tradeoff between notification latency detection accuracy; additional traffic does not always improve latency; Sympathy's process localization reduces.