Vodafone Works with Esri to Build UK-Wide Digital Twin on AWS

Technology communications company Vodafone, working with the geographic information system (GIS) company and AWS Partner Esri, built a nationwide digital twin for its UK mobile network hosted on Amazon Web Services (AWS). It is the first UK national digital twin to be built to support telecom infrastructure deployment. The platform is designed to improve Vodafone’s nationwide rollout of 5G cellular service and will enhance other services in years to come.

Vodafone, the leading pan-European and African technology communications company, has long relied on the geospatial software company Esri’s geographic information system (GIS) to make data-driven decisions on where to locate cell towers and how to improve its mobile coverage. When it decided to move from an on-premises GIS system to a cloud-based Esri ArcGIS Enterprise platform, it looked at the growing volume of data it had and realized it could gain even greater insights by building a full digital twin of its nationwide network.

It undertook the project with AWS Partner Esri, NTT Data, and ROK Technologies. ROK, which is also an Esri Gold partner and Esri Cloud Specialty Partner, specializes in architecting, deploying, and managing ArcGIS Enterprise and GIS solutions in the cloud.

Working with Esri, NTT Data, and ROK Technologies, and using AWS, a special projects team at Vodafone developed a location intelligence platform for understanding what its UK network looked like—from the location of cell towers to the height and canopy width of trees across the country. Now live and hosted on AWS, the system is helping Vodafone to optimize its UK network rollout and enhance other services to customers.

Vodafone’s service covers much of the UK’s population and depends on thousands of cell towers located across the country. Maintaining that network—and improving it through equipment upgrades and new 5G capabilities—requires Vodafone to understand the surroundings of every tower and any possible future changes. Everything—from fast-growing trees that could interfere with coverage, to new building developments bringing hundreds of new users to an area—must be considered.

As it pulled together the data needed to move from an on-premises GIS system to the ArcGIS Enterprise platform on AWS, Vodafone’s GIS Strategy team first envisioned using that information to enhance the company’s existing 2D network map with a 3D version. From the outset, the team understood that the large amount of data it had could form the foundations of a digital twin—a virtual representation of its cell towers and the surrounding landscape—so it proceeded to develop a proof of concept.

From the start, this was an ambitious project. Vodafone wanted a nationwide digital twin for the UK, including 3D buildings, trees, and other relevant features, which amounts to a huge volume of data that needs to be rendered and analyzed rapidly by the GIS software. “Everything we wanted to do was national (245,000 square kilometers),” says Boris Pitchforth, lead architect at Vodafone. “That makes the datasets so large in terms of the scale of what you’re trying to do—modeling millions of objects and billions of rows of data, things can get quite unwieldy. Using AWS, we can manage all of that data quickly and easily. We don’t have a limitation.”

Credit: Vodafone

For its proof of concept, Vodafone thought that it could limit the big data challenges by focusing on just a small region of the country to start. The team began by focusing on central London but, even then, the volume of data required was huge.

Amazon Simple Storage Solution (Amazon S3) provides object storage built to retrieve any amount of data from anywhere. “AWS could provide the scale and flexibility that Vodafone needed,” Pitchforth says. But before work could begin, Esri’s ArcGIS Enterprise platform first had to be architected and deployed in AWS. Because ROK Technologies is a partner of both AWS and Esri, it was able to provide valuable insight into this process. “ROK’s core expertise lies with the deployment of Esri’s ArcGIS Enterprise in AWS,” says Jason Harris, chief technology officer at ROK Technologies. “We are able to leverage advanced scalability and resiliency while taking advantage of cloud-native services that integrate deeply with the Esri stack.”

After the ArcGIS Enterprise platform configuration was optimized on AWS, the Vodafone team began building the proof of concept. The GIS provides a standard for all the different formats of data involved, reducing the risk of inaccuracies. A lot of experimentation was still involved in bringing all the data together so that it could be properly rendered to enable high performance.

Multiple resolutions of data have been integrated into this model, including detailed high resolution, so data can account for features such as narrow rivers or train tracks that could affect the placement of cell towers and, consequently, network performance. Working together, the teams were able to incorporate all of the data, creating a high-performing platform that can be adapted to meet changing conditions.

Working at that scale required a lot of thought about the mix of AWS services that would best handle such large data volumes. “We had to rethink how we were actually going to load it,” says Pitchforth. The team started with a foundation built on Amazon Elastic Compute Cloud (Amazon EC2), which provides secure and resizable compute capacity for virtually any workload, and Amazon Relational Database Service (Amazon RDS), a collection of managed services that makes it simple to set up, operate, and scale databases in the cloud.

In some instances, the team found that the best solution for managing the large volumes of data meant using Amazon S3. For other data management needs, it chose to use Amazon FSx—which makes it easy and cost effective to launch, run, and scale feature-rich, high-performance file systems in the cloud. The team is working to further enhance the platform with a hybrid Kubernetes deployment.

Whatever the challenge, the team was committed to finding a solution, and knew that AWS could support a high-resolution digital twin that also loaded quickly. “Never was it a thought that we can’t load it quickly enough or we have to go to a simpler resolution,” says Pitchforth.

By the time the project was ready to roll out, the team had a digital twin that represented 40 million geographic features, over half a million network features, and billions of rows of network performance data. “Everything needs to be able to cope with a large number of spatial features and their attributes, for both rendering and analysis,” says Dr. Rebecca King, former geodata scientist at Vodafone.

The scale of the data required put the project at “the cutting edge” for Esri’s technology, says Richard Stevenson, former telecom account manager at Esri UK. “We had different terrain models, building layers, vegetation, and all the infrastructure that Vodafone had. It was a very challenging project.”

At the same time, it was important to make the digital twin user friendly and efficient. “Optimizing the model was crucial to user experience,” says Alvaro Silgado, head of GIS at NTT DATA Spain, which helped with data loading, optimization, visualization, and integration. “We made use of the i3S GIS data container specification to create a ready-to-use cache with all the buildings and vegetation that could be stored directly in an Amazon S3 bucket. Portal for ArcGIS can directly consume the data from the Amazon S3 bucket, avoiding server usage and making a faster service for such a large amount of data.”

The digital twin has been available since 2021, and the initial target audience was engineers, who are responsible for determining cell tower placement. “It gives them greater insight,” says Pitchforth. “If they’ve got 3D visualizations of all the parameters, they can visualize it very quickly and rapidly assess the impact of environmental factors, such as tree growth, on the network elements.”

Vodafone is now exploring taking the digital twin concept to other markets and integrating data from other organizations—such as utility companies and local councils—to provide even richer location details. Other enhancements include adding machine learning for greater insights and automation.

As a “living thing,” the digital twin will be refined and updated over time to ensure it continues to reflect the real-life circumstances across Vodafone’s networks. “It’s always going to evolve,” Pitchforth says. “We’ve always known from the very beginning that we will keep fine-tuning and expanding this. As technology keeps advancing, we’ll be able to grow the digital twin’s capabilities so we can keep reaping more and more benefits from it.”