Where can CSPs use digital twins? Everywhere.

In simplest terms, a digital twin is a digital replica of a physical object or system. Today’s digital twins are powerful tools for innovation, but CSPs haven’t known where to start using them. That’s changing with the emergence of clear use cases for every part of a CSP’s organization, from internal operations to partner ecosystems. CSPs who seize the opportunity will reap the benefits of what Accenture calls the mirrored world — a virtualized reflection of reality that opens “a cornucopia of new opportunities and ways of doing business”.

Digital twins started out as virtualized representations of physical things. In many cases today that’s still how they’re used, for everything from aircraft engine design to building systems for monitoring and optimization.

But more and more, digital twins are also being applied to systems, processes, behaviors and even other digital phenomena. A digital twin can be essentially an abstraction of any “thing” used to achieve a goal or outcome based on the analysis of a purpose-defined data set. Going by this broader, more versatile definition, many CSPs could find they’re already using digital twins without realizing it — to manage customer relationships, for example, or deliver customer care.

A convergence of innovations has unlocked enhanced digital twin capabilities: cloud-based data capture, data processing, data analytics, simulation and artificial intelligence (AI)-based decision-making as well as increasingly sophisticated network virtualization and machine-readable network element configurations.

“The sheer power of the cloud and analytics means that we can consider simulating things we may not have simulated in the past,” says Appledore Research Group Principal Analyst Francis Haysom.

As a result, digital twins are catching on across the full spectrum of industries. Appledore Research Group projects the digital twin market to hit US$10 billion by 2025. Accenture reports that 18 percent of companies are scaling up digital twins in 2021 and 24 percent are experimenting with them.

The possible use cases for digital twins are now constrained only by a solution developer’s creativity, the availability of raw data and suitable analytics to construct a model. For CSPs, there are four areas of focus where digital twins have particular potential to yield meaningful business results: operations, processes, customer experience and partner ecosystems.

“I think as we look into the future, having virtual models in a mirrored world that integrates ‘intelligent digital thoughts’ will drive the next-generation innovation we're looking for,” says Tejas Rao, Managing Director and Global 5G Offering Lead for Accenture’s Network Practice.

Radio network asset planning is a good example of how digital twins can bring value to CSP operations because it’s an arena where there are no perfect decisions. Antennae can’t always be located in the absolutely best places and optimal coverage is seldom affordable. The goal is to field the best network coverage possible within a defined capital budget, so the more combinations and permutations a CSP can test before deployment, the greater the odds of finding the truly “best possible” configuration.

Digital twins can produce a virtual model of the network and support this kind of “what if” analysis, identifying and resolving potential gaps between coverage and demand. It requires creating an accurate digital representation of the radio access network (RAN) and its current demand, and comparing that demand against existing coverage. Using a user traffic and coverage map digital twin, one CSP in London, England, was able to improve average cell throughput by 26 percent and double their cell edge performance.

With the right forecasting algorithms, network planners can weigh the pros and cons of decisions such as whether to upgrade an existing antenna to massive MIMO or add a small cell to boost capacity in a specific location before incurring the expense of purchasing and commissioning new equipment. If they add even more information into their data set, such as antenna types and locations, city maps and building locations, CSPs can determine potential gaps between exterior and in-building coverage.

One Asia-based CSP that incorporated this kind of data was able to assess disparities between exterior and in-building coverage, boosting the return on investment (ROI) potential of its enhanced mobile broadband (eMBB) use case by 55 percent.

Digital twins also have a role to play in network automation — for example, optimizing RAN operating configurations for physical cell identifiers and automatic neighbor relations. If these are configured improperly, they can lead to call drops, slow call setups or tower-to-tower hand-overs that have a negative impact on the user experience. A digital twin in the form of a self-organizing network (SON) can enable analysis and corrections that can be pushed into the network for continuous monitoring and optimization.

These dynamic capabilities bring the whole concept of “planning” closer to live operations. “Something that would have been seen as a planning function by a telco in the past can now be considered as an ongoing network management function,” says Haysom from Appledore Research Group.

“Something that would have been seen as a planning function can now be considered as an ongoing network management function.”

Francis Haysom, Appledore Research Group

Digital twins can be used to develop, hone and train teams on managing network infrastructure or creating and deploying 5G services — any process that can be realistically simulated and experimented with. Today numerous companies worldwide specialize in building three-dimensional virtual spaces that workers can “enter” and interact with via virtual reality equipment to practice tasks without risk of consequence.

Cell site deployment is a repetitive process that benefits from digital “twinning” because of how diverse the variables tends to be in the real world. Cell sites are highly complex environments with common components but are never exactly the same configuration. Every aspect of any new installation, from the antenna to its mounting hardware, cabling, baseband unit, backup batteries, fiber backhaul connectors and more all have to be precisely specified to ensure smooth deployment. There are network function configurations as well that need to be considered. It is a complicated and time-consuming task.

With a digital twin, CSPs can template the deployment process and then customize it for each site, creating individual virtual models with complete configuration details for review and certification. The twin can include a vast array of information, from part numbers and component quantities to drone footage of antenna mounting locations and conditions. Once certified, project managers can publish the cell site digital twin configuration details and specifications to tower engineers, radio network engineers, core network engineers, and optimization systems, streamlining and speeding the process end-to-end.

Digital twins can also be used as references for new products. “Rather than just simply say here's a new product, here's a new capability, and putting it out to the market, you can assess the impact after launch in light of new information and find ways to make the product better in real time instead of reacting six months after the product has been put into place,” Haysom says.

Mapping network performance to customer behavior

CSPs may already be using digital twins without thinking of it in those terms when it comes to customer experience. Notions of “putting the customer first” have been a focus for years. Early solutions used network conditions as a proxy measure of the customer experience, but over time it became clear that customer-affecting issues don’t always originate in the network.

Is the service operating properly? Are the customer’s applications performing as they should? Does a customer with a complaint already have open issues with the support team? These too have emerged as critical factors in measuring a customer’s experience. A well-constructed digital twin in the form of a customer experience index can answer them in near real time, enabling significantly improved customer interactions, stronger loyalty and less churn. By drawing data from different parts of the organization and the external environment, customer care managers can construct a picture of the overall customer experience of the network, customer care, marketing, billing and more.

This will have important applications in the 5G space, says Rao from Accenture. “As we think about some of the 5G use cases that are still in early infancy, how do we create digital models that let us understand the different ways people are generating traffic, all of the various interactions. The real-time collection of data will not only unlock the experience of how we want to drive these new use cases but also let us model the actual cost benefit analysis associated with that.”

Digital twins don’t pertain only to internal activities: they can also be tools for developing market opportunities by allowing partners in an ecosystem to experiment with new configurations and services. Vendors can use ecosystem-oriented digital twins to evaluate the performance of their applications in a simulated network environment. And digital twins can help ecosystem partners collaborate to prove a solution works or to resolve a particular problem. They provide, in essence, a sandbox for ecosystem partners to accelerate service innovation, ultimately paving the way for faster, easier rollouts of physical networks and software.

“When you think about a 5G use case like a smart manufacturing plant, you can easily picture an ecosystem that brings together a network equipment company with a systems integrator with a CSP to connect all the real-time sensors in the 5G network with the manufacturer itself to create an end-to-end supply chain in real time,” Rao says.

“As we think about some of the 5G use cases that are still in early infancy, how do we define the network connectivity we’re going to need?”

Tejas Rao, Accenture