Overview
A cross‑domain optimization solution based on a federation of Network Digital Twins through coordinated RAN and edge‑computing optimization, reducing energy costs while preserving service quality.
Our result
This Key Exploitable Result delivers a functional cross‑domain optimization methodology that reduces energy consumption in 6G networks by coordinating resource allocation decisions across both the radio access network (RAN) and the computing domain. The approach builds on the 6G‑TWIN vision of Network Digital Twins, using predictive models of traffic, load distribution, energy consumption, and computational demand to determine energy‑optimal configurations across multiple layers of the network. The result provides a joint RAN-Edge optimization framework capable of predictive orchestration of multi-band RAN, as well as computing workloads in response to potential optimizations made at the RAN level. The resulting methodology enables operators and researchers to evaluate how energy‑saving actions in the RAN affect computing resources, and vice‑versa, ensuring end‑to‑end energy reduction with minimal impact on service quality.
Needs and target audiences
Technologies and methodologies are needed to enhance imec’s networked systems portfolio, with focus on a library of energy optimization xApps deployable across diverse network environments. The project will deliver Network Digital Twins (NDTs) for real-time evaluation of energy-efficient resource management, supporting scalable energy-saving applications. It will modernize University of Antwerp curricula for 6G skills and enable dissemination via journals and conferences, strengthening imec and Accelleran solutions and fostering academic–industry collaboration.
Current stage
The current stage of this KER is at the development of isolated components that later compose the full end-to-end solution. At the RAN side, this KER leverages the AI RSG product (VIAVI), and the dRAX and the data telemetry collection from Accelleran, while at the computing side, data-driven models are being developed and tested in small lab setups.
Unique value proposition
A cross‑domain energy‑optimization methodology that manages edge‑computing resources in response to optimizations made at RAN level, leveraging NDT technologies.
Replicability
To ensure replicability of our results, we are open-sourcing some of the datasets, trained models, and code required to reproduce the experiments. Detailed instructions, environment specifications, and preprocessing steps will be provided, allowing any researcher to independently verify the results.
Keywords
- federated simulation
- ai-native architecture
- network digital twin
- network planning
- management and control
Research project acronyms
- 6G-TWIN
Research projects
- Integrating Network Digital Twinning into Future AI-based 6G Systems
Project identifiers
- 101136314
Contributors
- ACCELLERAN
- INTERUNIVERSITAIR MICRO-ELECTRONICA CENTRUM
Source: Horizon Results Platform