A Formal Foundation for Failure Avoidance and Diagnosis

Kelly, Terence ; Wang, Yin ; Lafortune, Stephane ; Welsh, Matt

HP Development Company

关键词: fault tolerance;    failure avoidance;    concurrent programming;    sensor networks;    discrete control theory;   

RP-ID  :  HPL-2009-203

学科分类：计算机科学（综合）

美国|英语

来源: HP Labs

【 摘 要 】

This paper argues that Discrete Control Theory (DCT) provides a useful formal foundation for failure avoidance and diagnosis in a wide variety of computing systems. Our experience applying DCT to several difficult systems problems during the past three years convinces us that this powerful, general, mature, and rigorous body of theory belongs in the standard dependability toolbox. It is particularly valuable in new contexts thrust upon us by recent technology trends, including sensor networks and the multicore revolution.