The Role of Monitoring Frequency in Optimizing Digital Twin Networks

Abstract

Digital Twin Networks have garnered increasing interest in both industry and academia for managing, monitoring, and controlling networks, irrespective of the underlying technology. However, in networks characterized by low frequency and throughput, the monitoring data is often inadequate to provide sufficient insights into network behavior on its own. In essence, establishing bi-directional communication between the Digital Twin Network and the physical network is crucial for gaining insights into the network and enabling informed reconfiguration based on those insights. This is ideally achieved without relying on out-of-band communication. Therefore, alternative paradigms must be considered. This study investigates the impact of monitoring frequency on the reliability and accuracy of a Digital Twin Network within a Bluetooth Mesh environment. Multiple frequency settings were tested across two experiments to evaluate their effect on network performance. The comparison between real network paths and Digital Twin Network simulations provides insights into its behavior. Results indicate that higher update frequencies generally improve Digital Twin Network performance by enhancing path prediction accuracy. However, this improvement is counterbalanced by a reduction in network reliability at higher frequencies. Furthermore, it was noted that while higher frequencies tend to result in more accurate paths predicted by the Digital Twin Network, they also adversely affect the accuracy of end-to-end latency predictions. These findings highlight the need for a strategic balance between requirements in Digital Twin Network deployment.