Abstract: There are a large number of information system equipments in data center. It is so often that fault of single device hardware or software occurs. In order to protect the business continued service, the rapid recovery and automatic processing of the fault is necessary.In this paper, we mainly study the automatic recovery technology of the information system, and use the BP neural network to evaluate the software and hardware status and to predict the system capacity. An automatic disposal script set is constructed. Combined with the problem processing knowledge base on system status and script matching, the automatic recovery of the typical problems of information system is realized. From our experiments, the proposed method is feasible and can be used in data center.

Key words: power system, information system, fault, BP neural network, automatic recovery, openation and maintenance efficiency