The Research Of Fault Diagnosis Based On Communication Network Log

With the rapid development of modern information technology, the enterprise and the individual need a more safe and reliable telecommunication network system. So it is critical important to achieve fast network fault diagnose methods. And using the log information generated in the operation of the communication network has become a hot topic in the research of network fault diagnosis in recent years.This paper is based on the project which provided by the partner company. We first set active monitoring data, then extract them from the whole log information collection. The network fault method we proposed include two aspects:in the form of time dimension, we mine the association rules from alarms and generate frequent alarm sequences; in the form of spatial dimension, we use fault tree analysis locate the root cause node for fault in network elements. Our research works include the following three specific points:The first point, through the study of traditional Apriori, WINEPI algorithm and analyze their numbers of scan as well as storage structure, this paper proposes an novel sequential pattern mining algorithm called Prefix tree-WINEPI based on Prefix-tree. We design the contrast experiment and the experimental results verify that the improved algorithm is more efficient in terms of execution time and incremental processing than WINEPI algorithm.The second point, according to the research of log information characteristics, we set up the active monitoring items and use the formula to calculate them for rapid extraction from the whole log databases. That can reduce the interference from irrelevant information. At the same time we put forward a comprehensive diagnosis method which combines alarm correlation analysis in time dimension with fault tree analysis in space topology. It is applied to network fault diagnosis.The third point, we translate the theoretical work into practice. Through diagnostic tool design document writing, algorithm design and data storage design works, we implement a specific log diagnose tool which using Java Web programming and the relevant technology of the Web front-end. Through the fault analysis of the log information, this paper solves the problem of fault diagnosis successfully.In a conclusion, this paper combines the theory research and specific log diagnostic practices, which verify the effectiveness of the proposed methods and solve the log diagnose problem in the scientific research project.