Failure Diagnosis for Configuration Problem in Storage System

Configuration problem in storage system is one of the well recognized issues in today’s enterprise environment. Traditional approaches usually reply on deep knowledge of the underlying system. However, such solution usually cannot scale up when the system become large and complex. Recent research work has started using machine learning techniques to improve the overall performance of the diagnosis process. In this project, we try to identify the proper procedure to transform a configuration problem into a machine learning problem, which lays down the foundation of our future work in the area of configuration and change management.