DEPS: A Demand-Oriented Framework for Edge Intelligent Production System

In many cases, demands of Intelligent Manufacturing are generated locally and served locally, not necessarily dependent on the cloud. We believe in the distributed intelligent production system, cloud and edge systems should act as equal processing nodes for service migration and cooperation. This paper focuses on the edge nodes and expands the scope of edge computing, aimed at building an independent and complete edge system - DEPS based on demand-orientation. To adapt to the edge environment, DEPS adopts a hierarchical structure to provide a relatively complete system framework for multiple parallel services. To meet the demands of industrial production, DEPS takes two measures: 1) integrates a variety of existing industrial control systems; 2) uses flexible modular design to allow experts to participate in the design, enhancing the professionalism of the system which can be designed more on-demand. We use short and continuous industrial production data to evaluate DEPS and propose a method to measure its processing performance versus current systems: 1) as for a cloud only system, DEPS has superior performance servicing local demands in real-time, securely and of lower cost; 2) as for current edge systems, partial comparison of them in applications and performances is discussed.