Seamless Computing for Industrial Systems Spanning Cloud and Edge

Abstract

Industrial systems are more and more dominated by software. In addition, this software is present across several compute domains, from decentralized edge to centralized datacenters and clouds. While the productivity of application lifecycle management has tremendously improved in the area of cloud computing, there is no homogeneous and seamless environment to build, deploy and operate the software across these domains. This leads to a separation, inefficient processes, and duplicate efforts to provide software that runs across the different layers. This paper introduces the concept of seamless computing, which provides a seamless computing environment for multi-domain applications, supporting the mobility of workloads between cloud and edge. The approach is based on transferring established, de-facto standard cloud computing technologies to resource-constrained compute environments in the edge. After defining high-level requirements for seamless computing, a functional reference model is derived, and existing cloud technologies are discussed and selected according to their related capabilities.