Network Failure Detection Method Based on Deep Learning with Logs

With the rapid development of the Internet, the network architecture is becoming more and more complex and the number of network devices is increasing, which brings new challenges to network management and network security defense. In this paper, we propose a network failure detection method based on deep learning. Firstly, combining with the idea of natural language, the original network logs are represented in multiple dimensions such as words, events, and event sets. Secondly, the DBSCAN clustering method is used to analyze the similarity of log texts and achieve aggregation of logs of the same category. Then, after vectorizing the logs of the same category, network failures are detected based on a convolutional neural network. The experimental results show that the method proposed in this paper can effectively detect network failures with an average accuracy rate of over 96%.