Software Defined Networking in Advanced Manufacturing Systems

摘要

Robots have been widely used in industry to reduce the risk of severe injuries to human operators. Robots can operate using either local or remote (in the Cloud) computing re-sources. The latter case has several advantages over the former, including resource sharing, cost effectiveness, ease of maintenance, and scalability. In a typical scenario, data sensed by or for the robot need to be quickly and reliably transferred to one or more remote servers for further processing. Particularly important to the application is to rely on a reliable connection between the robot/sensor and the remote server.;In order to guarantee uninterrupted data transmission even in the presence of network out-ages, the development and adoption of new technologies such as high-speed flexible networks employing software-defined networking (SDN) technology are of the essence to flexibly interconnect the production floor to the cloud sites.;Firstly, we studied and found that when the CPU, memory, and network bandwidth resources that are allocated in the two environments (LAN vs WAN) are comparable, the execution time of the Godel software in the LAN and WAN environment does not differ significantly. Secondly, we proposed the PROnet Orchestrator, a SDN-based two-layer orchestrator software platform. The PROnet Orchestrator is designed to provision and maintenance end-to-end Ethernet flows through optical network. Finally, we ran industrial robot application over PROnet, and the experimental results demonstrated that the self-healing and prompt response of the Ethernet layer prevent a ROS-I application from being disrupted.