Vodafone and Cirrus360, a US-based tech company, have successfully tested AI-driven predictive digital twin technology that allows engineers to look into the future to see how compute for a mobile mast will perform even before they have installed it or introduced an upgrade at scale.
Predict outcomes
With AI reasoning applied to a predictive digital twin, engineers are given a dynamic virtual model of a physical 5G RAN site. They can then simulate future performance and predict outcomes before making costly upgrades or coverage improvements, saving both time and money.
Cirrus360’s Declarative Digital Twin powered Gabriel™ platform forecasts potential hardware or system failures such as memory, cache limits and Central Processing Unit (CPU) usage. It also predicts customer traffic and likely latency at a site before upgrading or deploying a new RAN. The system offers ways to optimise hardware and fine-tune software in advance, allowing everything to work seamlessly with silicon chips and other equipment. This approach is especially useful when implementing Open RAN which combines equipment from multiple vendors.
Vodafone anticipates Cirrus360’s Declarative Digital Twin will streamline hardware and software testing and optimise existing resources, such as detecting underused CPUs. This will allow engineers to concentrate on key tasks, reduce time spent in the lab, and accelerate network deployment. Typically, engineers will have to test between six and eight different RAN configurations before installing or upgrading a mobile network across a country.
Unlike traditional, rigidly programmed digital twins, a declarative digital twin uses high-level, rule-based descriptions (declarative specifications) to automatically adapt its structure and behaviour to a physical assets’ lifecycle. It defines what the system should do, not how, enabling automatic reconfiguration as the physical asset shifts from design to testing to operation. It also ensures boundary rules are followed, boosting accuracy and reliability.
Joint tests
The joint tests conducted by Vodafone and Cirrus360 in the UK and the US show how an open collaborative ecosystem can stimulate greater innovation. Vodafone is keen to extend this partnership to include other vendors through access to its real-world test environments such as the ones in Malaga, Spain, and Newbury, the UK, with support from new AI powered systems such as Cirrus360’s Gabriel™ platform.
Kyriakos Exadaktylos, Head of Network Architecture and Open RAN at Vodafone, said: “We are constructing a framework that will expedite testing and deployment of new innovations in purpose-built silicon and AI in the RAN. The Declarative Digital Twin using AI Reasoning has the potential to accelerate this process further and give our customers a greater level of service.”
“Our Declarative Digital Twin methodology, in conjunction with AI Reasoning in Cirrus360’s cloud-based Gabriel™ platform, acts as a RAN testing specific AI assistant for system integrators,” added Chaitali Sengupta, co-founder and CEO at Cirrus360.
Cirrus360’s Declarative Digital Twin platform was developed with funding from the US Department of Commerce, National Telecommunications and Information Administration (NTIA), and the Public Wireless Supply Chain Innovation Fund (PWSCIF).
