Application-Driven Bandwidth Guarantees in Collaborative Mobile Edge Computing

This paper proposes a novel bandwidth management framework for providing specific applications with bandwidth provisioning in a collaborative Mobile Edge Computing (MEC) architecture. We define and analyze the problem of bandwidth provisioning, which involves determining the optimal relationship between the bandwidth guarantees specified in Service Level Agreements (SLAs) and the combined bandwidth contributed by cloud and edge servers. Our objective is to minimize the bandwidth costs for the service provider while meeting the necessary SLA constraints. We introduce a new bandwidth provisioning framework called the little $\mathcal{R}$ framework, which is based on the fluid model. This framework efficiently satisfies the bandwidth requirements specified in application-driven SLAs while considering high availability. Through modeling and verification, we evaluate the impact of the framework on SLA bandwidth guarantees and delivery latency.