Evaluating Optimistic Synchronization of Network Graphs in Mobile Networks with MEC Support

Future networks with well-developed Multi-Access Edge Computing (MEC) resources, where network nodes and distributed service instances can form a dense topology, face the challenge of distributing service data among many nodes. In particular, frequent data changes require an efficient update mechanism. System flexibility and user mobility, e.g, cellular hand-overs, or unpredictable service deployment or node changes, pose additional challenges. In this paper, a method is proposed and evaluated, how network graph-structured data can be distributed and synchronized between many nodes. The design considers especially the challenges of mobile networks with dynamic changes. At the example of a MEC scenario, the method is evaluated using a system-level simulation environment with fifth generation (5G) cellular communication, mobile as well as locally bound nodes, and dynamic service deployment changes. Network and computing resources are represented and shared using a graph. To gain deeper insight into the performance of our model, the time and bandwidth resources required to replicate graph updates to all clients in the 5G network are evaluated. Hereby, the impact of different background network traffic, of information-dependent graph update frequencies, and of the connectivity type can be shown. Finally, by implementing a practical example, the maintenance overhead for synchronizing the graph data to have a local information source is compared to the time required when querying remote information sources.