Can digital twins deliver for telecoms?

The use of digital twins is spreading beyond manufacturing and into areas such as healthcare, construction and smart cities. These data-driven simulations of the real world allow organizations to identify and solve problems before they happen in real life, and to optimize systems, processes, operations and prototypes. All this sounds potentially useful for communications service providers (CSPs), but can they realistically harness digital twins today?

A digital twin is a digital representation of a physical object, process or service that uses real world data to create simulations to predict how a product or process will perform, including its environmental impact. The benefits of being able to virtually analyze and model how a system, process or product or prototype will work in reality before rolling it out is making digital twins increasingly popular. In Asia, for example, IDC estimates that between 2021 and 2027 the number of new physical assets and processes that are modelled as digital twins will increase from 5% to 60%. However, when it comes to CSPs, the use of digital twins is still nascent. China Mobile, for example, says it is in the early phase of digital twin usage and only in specific areas of its business. “Maybe network operations will first benefit from digital twins,” according to Xiaowen Sun, a researcher with China Mobile Research Institute. “Our real network cannot achieve online decisions and almost all the network operations need human involvement," says Sun. "If…[we can] make decisions based on sufficient calculation and verification in the digital twin environment, it will make network operations more accurate and reliable.” John Davies, chief researcher at BT, also sees digital twins proving useful for CSPs’ network operations. “For us, key areas are those related to network operations – planning and building, monitoring, maintenance,” says Davies. BT is also exploring the use of digital twins for energy planning and capacity management and investigating the “modelling of IoT devices and smart meters to offer resilience, scalability and scenario planning for our IoT platform”, explains Davies. Closing the loop Ideally, using digital twin modelling and analytic capabilities to evaluate end-to-end network performance from radio access to transmission to the core, “will bring everyone onto the same page to make a closed-loop decision with low trial and error tolerance”, says Jun Zhu, Chief Architect, Senior Product Director, AsiaInfo. Closed-loop decision-making is part of what manufacturers seek from digital twins. A national telecoms network, however, is much more complex than a discrete manufacturing site. And the dispersion of network assets creates very real obstacles to the extensive use of digital twins to inform business decisions. “Every component in the mobile network needs to be modelled in a digital twin network and the interactions and network services – e.g. PDU sessions – all need to be implemented in a digital twin. This is a very huge task,” says China Mobile’s Sun. The further a CSP transforms its networks and operations the easier digital twin deployment promises to be. In particular, virtualized, software-based networks will create the necessary foundation for the extensive use of digital twins in designing, optimizing and maintaining networks. Sun makes the point that it will be essential to be able to model the network component and establish a real-time interaction between the physical network and its digital twin. “For the core network part, the network function virtualization is a good way to achieve this,” she says. “All the network functions are provided by vendors as software-based on a unified network infrastructure; this is to a certain extent the preliminary stage of a digital twin.” Once the network and computing are converged, “it will be much easier to build digital twins of networks” that can be used to make business decisions, says AsiaInfo’s Zhu. Using digital twins today BT’s Davies acknowledges that having dispersed network assets represents a challenge to CSPs looking to create digital twins. Yet it is still possible to make use of digital twin technology today, believes Davies, notably to make more efficient use of energy across networks, to develop more sustainable field operations and to improve network asset management. “Telcos’ digital transformation will help realize a step change in their compute and data capabilities, but they do not need to wait until everything is in place to take advantage of what digital twins can offer.” BT’s research is examining several issues, including the federation of digital twins and interoperability, as well as the underlying infrastructures needed for digital twin and metaverse applications, according to Davies. It is also building use cases to understand which technologies, standards and architectures are needed to secure devices and digital twin data; to see how AI can improve insight; to model outcomes; and to gain insight into how to better optimize resources. Ericsson is exploring a number of discrete opportunities for digital twins. Last year it announced a trial with Vodafone in the UK to show how a digital twin can assist with discrete network site upgrades. Drones equipped with a high-definition camera collect data from network sites, which is then made available in the cloud as a digital twin model. Network engineering teams are then able to base decisions on the data without everyone having to visit the site. Ericsson has also been working with Nvidia to build city-scale digital twins to help accurately simulate the interplay between 5G cells and the environment so that CSPs can maximize network performance and coverage in an area of a city. Ericsson uses a city planning engine to accurately represent not only physical structures but also the materials they are made from, which impacts radio propagation, according to Håkan Olofsson, Head of New Concepts at Development Area Networks, Ericsson. And even though CSPs are not making much use of digital twins today, this could change as 5G-based services become more complex, says Olofsson. With the 5G platform “we have slicing and quality of service separation to handle the different services on top of the same network” he says. “As that becomes the norm, I think it's going to be more important before launch to understand what's going to happen.” Nonetheless, for China Mobile’s Sun and AsiaInfo’s Zhu digital twins still promise to be “mostly a 6G technology”, that will depend on the development of greater network intelligence. But once intelligent networks are in place, Sun believes digital twin technology can “provide a completely consistent environment for network decision-making and operation verification, which is of great significance for assisting decision-making”.