Agradecimentos: The authors would like to thank the São Paulo Research Foundation (FAPESP), grants #2018/12447-3, #2018/196395, #2015/24494-8, #2018/23064-8, #2020/05121-4, and FAPEMIG, grant #APQ-02510-18. Simulations were performed on UBELIX (http://www.id.unibe.ch/hpc the HPC cluster at the University of Bern Ver menos

Abstract: Edge Computing architecture provides computing capacity closer to users to fulfill the requirements of Future Internet services and applications, such as low latency. However, as users move at the edge and connect to different access points, service instances have to be migrated in order to keep service levels constant. Considering this issue, we introduce a graphbased algorithm to distribute data over networks considering time cost budgets. We then use this algorithm to develop a mobility-aware latency-constrained solution to position userspecific service instances, i.e., services that use application state and user session data, called Data Covers Framework (DCF). We compare DCF with other approaches for service provisioning at the edge that use fixed hosts or active service placement considering user mobility. Our experiments show that DCF achieves similar or better performance in terms of the percentage of packets delivered under latency requirements for different application classes while reducing the number of service migrations by 21% and the service interruption time by 41% Ver menos