Multifractal Analysis of Round-Trip Time for Computer Network Status Assessment

Abstract

Continuous network monitoring is essential for detecting anomalies, misconfigurations, and cyber-attacks that degrade system performance. Although Round-Trip Time (RTT) is a fundamental metric for assessing network status, the complex and non-linear nature of network traffic often makes conventional statistical analyses insufficient. This paper proposes an advanced approach grounded in non-linear dynamics to evaluate network states, employing multifractal analysis to RTT time series. By examining the behavior of the Legendre spectrum across varying packet sequence lengths and packet sizes, the study identifies key multifractal features that distinguish different network conditions. Data acquisition and processing were implemented in Python using specialized network analysis libraries. In addition, regression analysis was performed to quantify the relationships between multifractal features and packet parameters. The proposed method provides deeper and more robust insight into underlying network dynamics than traditional linear models, making it particularly suitable for computer network monitoring and automation.