Issue link: https://iconnect007.uberflip.com/i/1379105
72 SMT007 MAGAZINE I JUNE 2021 implementation to monitor the healthiness of the edge servers/devices. Data security Data backup and archive service shall be implemented to back up the edge data into an enterprise data warehouse. e service can run asynchronously between different edge devices and edge servers. e enterprise data warehouse shall be the interaction point for other applications or users to consume and retrieve the data, minimizing unnecessary access to edge device and edge server. Data is encrypted at rest and in transit between edge devices and cloud. Cost Cost is not directly reflected in this archi- tecture diagram. It is more a function of the implementation options of the architecture decisions above. erefore, cost factors shall be considered when deciding the implemen- tation option for each of the blocks in the diagram. Hybrid Quality Inspection Framework While manufacturing companies are start- ing to deploy AI assisted quality inspection onto their manufacturing floors, it would take a considerable period of time to con- vert all their quality inspection operations to be AI-assisted. From a manufacturing oper- ation point of view, a hybrid quality inspec- tion framework is required to support both AI-assisted inspection and human cognition inspection concurrently on the manufacturing floor. is is critical for both risk mitigation of potential disruption from newly developed AI services, and for smooth transition from human inspection to AI-assisted inspection at organization level. Moreover, there would cer- tainly be some critical inspections that require sampling verification by human cognition to ensure there is no dri of AI accuracy. In addi- tion, although some quality inspection would remain with human cognition, manufacturing companies shall automate and digitalize their quality inspection in the area of image acqui- Figure 2: Edge computing architecture illustration.