Detecting removed attributes in the cyber system for smart manufacturing

According to the concept of edge computing, some service and data should be moved from the centralized data server to the data source for enhancing the computational efficiency especially for real-time applications, e.g., production monitoring. In the manufacturing industries, most of the data sources are equipment. Deploying the application programming interface (API) services near the data source is the common solution. However, keeping the consistency of the production configuration between API servers and the manufacturing data center with minimum resource is a critical problem in terms of the computational efficiency and the speed of response. In this paper, we propose the partition and pass (PnP) mechanism to synchronize the production line settings saved in the API servers with that in the data center. From the analysis results, we show that the approximation ratio is 2 compared with current solutions including the time-based approach and the counter-based approach. It means that the proposed PnP only needs one half computing cost to finish the same number of products with that required by current solutions. Based on the simulation results, the PnP only requires 31% computation time to finish the synchronization process which matches the analysis results. Moreover, we provide a case study to show that the proposed PnP captures more events than current solutions. In other words, applying PnP in the production line will increase the computation performance to synchronize the production configuration between API servers and data center in saving manufacturing data.