State Management Strategies for Stateful Digital Twin Migration in Industrial Scenarios

While the adoption of digital twins (DTs) in industrial scenarios has been widely investigated, their dynamic migration remains largely unexplored. DTs fundamentally differ from traditional cloud applications. Due to their cyber-physical nature, DTs are often stateful and heavily rely on contextual and historical data. As a result, migration strategies designed for traditional stateless cloud applications may not be suitable for stateful DTs. To the best of our knowledge, no prior work has addressed the challenge of stateful DT migration. To bridge this gap, this work identifies three state management strategies (state replication, state rebuilding, and state sharing), develops six corresponding implementations (DT API, storage relocation, hot start, cold start, storage rebinding, and distributed cache), and experimentally evaluates them in an industrial setting.