Minghui Wang, Jiangxuan Xie, Xinan Yang, Xiangqiao Ao, AI Research Institute, H3C Technology Co., Ltd, China
The network is very important to the normal operation of all aspects of society and economy, and the memory leak of network device is a software failure that seriously damages the stability of the system. Some common memory checking tools are not suitable for network devices that are running online, so the operation staff can only constantly monitor the memory usage and infer from experience, which has been proved to be inefficient and unreliable. In this paper we proposed a novel memory leak detection method for network devices based on Machine learning. It first eliminates the impact of large-scale resource table entries on the memory utilization. Then, by analyzing its monotonicity and computing the correlation coefficient with the memory leak sequence sets pre constructed by simulation, the memory leak fault can be found in time. The simulation experiments show that the scheme is computationally efficient and the precision rate is close to 100%, it works well in the actual network environment, and has excellent performance.
Memory leak, Resource table entry utilization, Correlation coefficient, Time Sequence monotonicity, Machine Learning.