Research On High - Order Interval Estimation Algorithm In Network Traffic Monitoring

Higher network bandwidth and high throughput capacity are important to the development direction of the current network, with network transmission rates continue to Increase large, high-speed network for the monitoring technology is constantly improving, The main application of the technology is currently sampling the sampling interval or time-based packet, then the mean and variance estimates based on sample results in order to complete the evaluation of network traffic. As the magnitude of network bandwidth to improve data transfer rates and effective throughput is increased significantly, significantly increasing the sampling frequency packets required the use of sampling in the same circumstances, also will lead to increase in system resource consumption; but used of the sampling time interval will result in the same time interval and the amount of data increases the probability of random transitions. Therefore, in order to obtain a relatively accurate estimate, the need to improve the sampling frequency, and increased resource consumption ratio. So, optimize conventional flow estimation method based on relatively few realize the sampling frequency and resource consumption on a higher estimation accuracy is a more meaningful research.To accommodate more traffic monitoring network of high-speed data transmission, on the basis of the sampling frequency reasonable adjustments on the proposed high-end and interval estimation algorithms for processing data through its low-level sampling, which uses low-level measured sample data high-end estimate of the interval information, and evaluate high-speed network traffic based on the estimated size of the high-end estimate of the interval information. The article proposed the KL divergence theory on the accuracy of network traffic monitoring for evaluation to packet sampling, time sampling, low-order sampling interval used to high-order interval estimation methods, On the basis of the monitoring programs on different flow system resource consumption and precision performance simulation comparison and description.Simulation results show that, the network bandwidth and data traffic is increasing, the proposed algorithm can reduce the number of sampling system, reducing system load and improve the accuracy of the flow monitoring to ensure the effectiveness of network traffic monitoring, At the same time as the sampling interval increases, the combined effect of high-level estimation algorithm is more obvious, on the one hand and effective solution to the problems of high latency and high resource consumption caused by higher sampling frequency, on the other hand flow estimation accuracy is significantly improved relative. It can be seen that the proposed method has better use value of the article in a more high-speed network traffic monitoring.