Network digital twins (DT) are becoming more popular among telcos worldwide as a way to address various business challenges — from sales and marketing, network operations and customer experience, to IT operations, network planning, and more. This piece explores practical applications of network DTs and how they’re empowering businesses to achieve their autonomous operation objectives.

The Power Behind Network DTs

A network digital twin is a large-scale model that mirrors the real physical network across its life cycle and is continuously updated from design, operational data, and sensor data (real network traffic data). It can use simulation, machine learning, and reasoning to help make decisions or check the decisions made by machines or network operators. It can also simulate what-if scenarios to apply the changes efficiently. DTs can improve the operational efficiency and network quality for telcos. Some key examples are:

  • Network Change Management: Network digital twins can be used to test new network configuration changes before they are deployed in the real world. This can help to identify and fix problems early in the process, saving time and money.
  • Network Performance Improvements: Network digital twins can be used to analyze network performance and predict the performance of various services and networks. It can identify the potential degradations and assist in preventive steps.
  • What-If Scenarios: Digital twins can be testbeds for assessing scenarios related to network planning to determine the best configuration. DTs can create what-if analysis of new networks. This can help to identify and fix problems early in the process, saving time and money.
  • Validating Machine Decisions in Autonomous Operations: DTs can help in validating the closed-loop automation decisions made by AI/ML models before decisions are implemented in the network. This will help in improving the trustworthiness of autonomous systems.
Key technologies that enable network digital twins include artificial intelligence/machine learning (AI/ML), modeling tools, simulators, graph technology, cloud computing, accelerated computing, virtually unlimited storage, model-based design (MBD) methods, 3D simulation, data engineering, and data science. NVIDIA Omniverse, a platform for developing Universal Scene Description (USD) applications for industrial digitalization, provides several important features that make it especially well-suited for building network digital twins.
With expertise in telco networks and extensive AI investments, Wipro is focused on creating a Wipro reference network digital twin on the NVIDIA AI platform, using NVIDIA Omniverse and NVIDIA Tensor Core GPUs.
The following diagram illustrates the architecture for a typical network digital twin and its interfaces to an ecosystem of applications with the operations support systems (OSS) of telcos.
Unleashing the Power of Network Digital Twins

Click here to view the architecture overview

Creating the reference digital twin for a 5G NR network involves examining and assessing the following components:

  • DT Foundation Model: A network DT foundational model is developed using the physical, virtual, and logical inventory of a network, then superimposes the network configuration, services, and environment data onto this inventory. The network DT model should accurately represent the physical network components, including routers, switches, firewalls, and servers. This involves capturing their configurations, capacities, and functionalities.
  • Data Analysis: Techniques like machine learning and anomaly detection are applied to extract insights from network data. These insights can then be used to update the network DT model and predict future network behavior.
  • Actionable Insights: Network digital twins analyze real-time data and provide actionable insights related to network performance degradation, predicting equipment failures, or suggesting network optimization strategies.

DT Applications and Their Benefits

Wipro has leveraged its technical expertise in these areas to develop several reference applications for network digital twins on NVIDIA Omniverse. Below are a few examples of these applications and their potential benefits.

5G NR Traffic Steering Application: In a multi-layered radio deployment of 5G NR with low-, mid- and high-band-based cells, it is important to balance the traffic across layers while maintaining a quality user experience. Traffic steering applications analyze various network parameters like cell congestion, signal strength, user data requirements, service type, and user mobility. Based on this analysis, the rApp/xApp provides inputs to the radio resource management layer to admit the user equipment to specific cells and frequency bands or component carriers, while balancing the load by moving traffic across layers.

Traffic steering rApps/xApps consider user-specific and service-specific factors, offering more intelligent and dynamic traffic management.

Self-learning AI techniques are used to measure and predict traffic conditions. A probabilistic policy recommendation is applied to balance the traffic on carriers and deployed in a geographical location. Wipro developed a reference application leveraging the NVIDIA AI platform with AI model training and inferencing run on NVIDIA Tensor Core GPUs, and digital twin 3D modeling run on NVIDIA Omniverse.

Unleashing the Power of Network Digital Twins

Key benefits:

  • Improved network capacity and user throughput
  • Reduced congestion and dropped calls
  • Enhanced user experience with better quality of service (QoS)
  • More efficient use of radio resources
5G NR Slice SLA Assurance: To meet B2B contract requirements, telcos need to ensure that RAN slices deliver on their SLA promises. Wipro’s reference application continuously monitors various performance data of each RAN slice, including throughput, latency, packet loss rate, jitter, successful access attempts, and handover success rate. The applications use an AI/ML model to predict SLA adherence of slices and suggest necessary steps to avoid SLA violations. Closed-loop automation adjusts resource allocation across slices and redistributes radio resources like power and bandwidth to the underperforming slice.
Self-learning AI techniques are used to measure and predict the slice assurance performance management, KPI, and SLA calculations. Whenever slice assurance performance is degraded, an AI-based control module generates an action to reconfigure the slice executed. The application uses NVIDIA accelerated computing to train the models and to perform inference based on live network data.
Key benefits:
  • Optimize network resource utilization by dynamically adjusting resources across slices
  • Reliable and predictable performance for different network services running on separate slices

Wipro also built a reference network digital twin on the NVIDIA Omniverse platform, using a layered architecture and AI-based libraries for network automation, adhering to 3GPP, O-RAN, and Tele Management Forum (TMF) standards.

“As telcos embrace the future of connectivity, network digital twins will play an indispensable role in shaping the telecommunications industry and beyond,” says Lilac Ilan, Global Head BD Telco, AI Powered Operations, NVIDIA. “They enable telcos to enhance performance, optimize control, and leverage decisional intelligence through AI automation. Combined with LLM, this will provide telcos a faster, more dynamic way to interact, test and operate the network.”
Thomas Mueller, CTO, Wipro Engineering Edge, Wipro notes that “Wipro is at the forefront of this industry landscape change with the network digital twin solutions built on the NVIDIA Omniverse platform, using a tiered architecture and AI-based libraries for network automation, following TMF standards.”
With continuous synchronization between physical and virtual twins, Wipro’s network digital twin solution provides a comprehensive overview of the network ecosystem, transforming businesses by accelerating holistic understanding and optimal decision-making.

About The Author

Ravi Kumar Emani

General Manager and Practice Head, 5G Edge Connectivity, Wipro Engineering

Ravi has more than 25 years of experience helping global enterprises realize their connectivity goals. He is currently responsible for the Connectivity Practice Unit for NEPS and the Communications portfolio for Wipro Engineering. Ravi has authored numerous articles on 5G and is a Distinguished Member of the Technical Staff (DMTS) at Wipro.