A Service Placement Algorithm Based on Merkle Tree in MEC Systems Assisted by Digital Twin Networks

This paper considers a digital twin network (DTN) assisted mobile edge computing (MEC) system. When wireless device (WD) users request services from MECs, we assume in the system there are a cloud server, storing all of service entities, and a lightweight digital twin network, providing the digital replicas of state information of all MECs instead of service entities. We aim to maximize the number of service requests served by the MECs, or, equivalently, to minimize the load of the cloud. This is formulated as a mixed-integer non-convex optimization problem. We propose a service placement algorithm based on hash-based data structure called Merkle tree to solve the problem. The introduction of candidate mode pruning effectively reduces the time complexity of the algorithm in iterations. Simulation results show that our proposed method has a better performance compared with the other benchmarks.