A New Virtual Network Topology based Digital Twin for Spatial-Temporal Load-Balanced User Association in 6G HetNets

• Published in: IEEE journal on selected areas in communications Vol. 41; no. 10; p. 1
• Main Authors: Jia, Pengyi, Wang, Xianbin
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.10.2023; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: 6G mobile communication; Communications traffic; Complexity; Complexity theory; digital twin; Digital twins; Dynamic scheduling; Graph matching; Graph theory; heterogeneous networks; Load balancing; Network topologies; network topology; Optimization; Principal components analysis; Provisioning; Quality of service; Quality of service architectures; Radio equipment; Real-time systems; spatial-temporal analysis; user association; User requirements; Virtual networks; Virtual reality
• ISSN: 0733-8716; 1558-0008
• DOI: 10.1109/JSAC.2023.3310104

Dynamically associating distributed mobile users with proper base stations in 6G heterogeneous networks (HetNets) becomes critical to achieve both diverse quality of service (QoS) requirements of all users and entire network performance. However, the significantly increased complexity of matching the irregularly distributed users and base stations as well as highly dynamic network traffic often cause unbalanced spatial-temporal loads for multi-tier base stations during user association. To overcome this challenge, we propose a new virtual network topology-based digital twin to reduce the complexity of load-balanced user association in 6G HetNets. During the digital twin construction stage, instead of using highly dynamic low-level physical layer attributes (e.g., channel conditions and SINR), we intentionally consider more stable and relevant communication performance indicators and physical statistics to effectively reflect both real-time link quality and overall network dynamics. To assist overall network operation, fast update of the digital twin for HetNets is achieved by adopting principal component analysis to discover specific network areas with changes. To improve the overall QoS provisioning and network performance, the proposed virtual topology-based digital twin is further utilized to predict the spatial-temporal dynamics of HetNets for more balanced user association by bipartite graph matching. Simulation results show that the proposed method can construct effective digital twins and support load-balanced user association with maximized network-wide QoS satisfaction.