Safety Analysis of Rail Transit Redundant Structure in Cloud Computing Environment Based on Graph and Bayesian Theory

The migration of train control systems to cloud computing is an important trend of urban rail transit. A traditional redundant structure (T-RS) is a safety related vital part of the train control system. In this paper, the safety of the redundant structure in cloud computing environment (CC-RS) is researched. Firstly, a deployment scheme of the CC-RS is presented, the types of failures, especially those related to cloud computing, are analyzed. Then, a model is setup based on graph theory which takes the physical nodes, virtual machines and fault-tolerant mechanisms of cloud computing into consideration. After that, the safety of the redundant structure is analyzed using a minimal spanning tree method and Bayesian algorithm. Lastly, some optimization measures on the deployment of the redundant structure are proposed and verified through numerical analysis, according to which the redundant structure in cloud computing environment can be safer than the traditional redundant structure.