Distributed Simulation for Digital Twins of Large-Scale Real-World DiffServ-Based Networks

TL;DR

This paper addresses the challenge of high-fidelity, scalable online simulation of DiffServ-based networks for Digital Twins. It introduces Quaint, an optimistic PDES toolkit built on the ROSS engine, paired with a novel high-efficiency event-handling model for DiffServ routers to drastically reduce simulation events. Through real-world metropolitan topology experiments, Quaint achieves up to 53.66x speedups in sequential mode and up to 232.88x in distributed mode compared with OMNeT++/INET, while maintaining accuracy in QoS metrics. The work also demonstrates the critical role of workload-aware partitioning for scalable distributed simulation and outlines directions for future QoS modelling and workload-aware partitioning strategies.

Abstract

Digital Twin technology facilitates the monitoring and online analysis of large-scale communication networks. Faster predictions of network performance thus become imperative, especially for analysing Quality of Service (QoS) parameters in large-scale city networks. Discrete Event Simulation (DES) is a standard network analysis technology, and can be further optimised with parallel and distributed execution for speedup, referred to as Parallel Discrete Event Simulation (PDES). However, modelling detailed QoS mechanisms such as DiffServ requires complex event handling for each network router, which can involve excessive simulation events. In addition, current PDES for network analysis mostly adopts conservative scheduling, which suffers from excessive global synchronisation to avoid causality problems. The performance analysis of optimistic PDES for real-world large-scale network topology and complex QoS mechanisms is still inadequate. To address these gaps, this paper proposes a simulation toolkit, Quaint, which leverages an optimistic PDES engine ROSS, for detailed modelling of DiffServ-based networks. A novel event-handling model for each network router is also proposed to significantly reduce the number of events in complex QoS modelling. Quaint has been evaluated using a real-world metropolitan-scale network topology with 5,000 routers/switches. Results show that compared to the conventional simulator OMNeT++/INET, even the sequential mode of Quaint can achieve a speedup of 53 times, and the distributed mode has a speedup of 232 times. Scalability characterisation is conducted to portray the efficiency of distributed execution, and the results indicate the future direction for workload-aware model partitioning.

Figures (6)

• Figure 1: The internal structure of a router under study. The number of ingress and egress ports can vary in different routers.
• Figure 2: The proposed event handling model of a router. The workflow starts from an event (a green ellipse) and ends at a sink or creating a new event. A "try to send" procedure is extracted from the main workflow and shown at the bottom.
• Figure 3: Topology of a real-world metropolitan-scale bearer network. Black dots are "access" routers. Blue dots are "mixed" routers. Red dots are "kernel" routers. An access router usually has 1 to 3 ports with 25Gbps bandwidth. A mixed or access switch can have over 10 ports with 10-100Gbps bandwidth.
• Figure 4: Comparison of the accuracy and efficiency between Quaint and INET
• Figure 5: The execution time of the proposed simulator given different levels of parallelisation (number of processes) using different partitioning approaches.