Reliable Resource Matching in 6G Computing Power Network for Digital Twin Service

Due to the complex and dynamic service demands involved in digital twin services, traditional resource matching methods often fail to meet the requirements. To address this challenge, we leverage Graph Neural Networks (GNN) for task graph matching to enhance the deployment of computing resources. The crux of our approach is an adaptive service decomposition process that breaks down complex services into smaller, more manageable subtasks. These subtasks are then strategically deployed across various nodes within the computing power network (CPN), utilizing distributed resources to maximize efficiency. Moreover, we employ dynamic task scheduling based on Service Level Agreement (SLA) constraints to prioritize tasks, thereby enhancing the quality of computing resource matching. The deployment strategy is grounded in GNN-facilitated low-complexity deep graph matching techniques, aimed at effectively integrating network functionalities with service demands. Our proposed service matching method based on GNN considering dynamic task scheduling (DTSM) has been validated through simulation experiments, with results indicating significant improvements in load balance and overall service processing delay.