A Digital Twin Framework for Edge Server Placement in Mobile Edge Computing
Elif Bozkaya
Abstract
Mobile edge computing is considered a key technology for upcoming 6G networks, which allows real-time processing of large amounts of data at the network edge. Specifically, considering the strict requirements of new services and Internet of Things (IoT) applications in different domains, such as healthcare, transportation, smart city, smart grid and more, increased mobile data traffic can be handled with the benefits of this technology. As an enabler for these interesting applications for 6G, mobile edge computing deploys edge servers at the network edge, enabling the offloading of computation tasks to reduce processing latency. However, achieving cost-effective placement of edge servers presents a challenge in effectively distributing the workload among them while simultaneously minimizing overall system cost. To address this issue, we propose a digital twin framework, where digital twin technology is used to model virtual replicas of edge servers in the real world and simulate their behavior. Then, the proposed edge server placement framework can place the edge servers and balance the workload of the edge servers. The objective of the proposed strategy is to minimize the system cost in terms of processing time and access delay. The edge server placement is formulated as an optimization problem in mobile edge computing systems and a genetic-based algorithm is introduced to address this issue. Simulation results demonstrate the efficiency of our proposed digital twin framework achieving less system cost compared to other benchmark methods.