A Cloud-Edge Computing Approach for Epidemic Close Contact Tracing Based on Spatial-Temporal Trajectory

Contact tracing, as an effective method to slow the spread of epidemics, is traditionally performed with manual surveys on positive cases to acquire information about close contacts, which induces an unbearably long tracing duration. In this paper, we propose a cloud-edge computing-aided contact tracing system where multi-access edge computing (MEC) servers are introduced to collaborate with the cloud center (CC) to mitigate the duration issue. Collecting the spatial-temporal data of mobile users, the MEC-integrated base stations organize these data via the R-tree structure for efficient tracing task processing. After aggregating spatial-temporal data from MEC servers, the CC partitions and assigns contact tracing tasks to different servers for parallel computing. An iterative transmission bandwidth allocation and task assignment algorithm is designed to minimize the overall tracing duration, where the correlation of the cases' positions and MEC servers as well as the limited buffer size of MEC servers are considered. Simulation results show that with the proposed scheme, the tracing duration is reduced by 55.8% compared with cloud computing, and a 100% tracing accuracy is achieved.