Fault detection of energy‐aware grid systems in big data environment

Abstract

SummaryGrid technology based on parallel theory is used to study a large amount of data processing. This research mainly discusses the fault detection of grid dynamic system in a big data environment. In order to verify the communication efficiency between different security nodes of the ZSS system under different operating system conditions, this study will use four physical computers and VM Ware software to establish a simulated communication environment, using GlobusToolkit (GT)'s existing deployment mechanism to realize the overall grid dynamic deployment architecture. Using Web service technology remotely calls the monitoring module interface of resource nodes in the domain and uses standard protocols to send monitoring data. At the same time, in order for the management node to save the monitoring data in the local standardized form of XML, the administrator can use the saved historical data to process system failures, analyze, and predict system performance. In order to solve the problem of information explosion, a fault detector is used to precisely control the cost of message detection and the state distribution of the detection result and the time of message delay. Under special circumstances, it takes 60 min to execute a 30‐min job, and the fault detection point consumes an average of 54.33 min. The research results show that the fault‐tolerant strategy can improve the performance of resource consumption in the fault‐tolerant tracking system. Adopting the method of dynamically adjustable detection interval can effectively reduce the overhead of the grid dynamic system. The detection method used in this study can meet the application requirements of the grid.