Internet-of-Things monitoring system of robot welding based on software defined networking

In view of the heterogeneous characteristics and enormous connections of the underlying networks in Internet of Things(IOT), we proposed a SDN-IoT architecture including physical layer, control layer and application layer in this paper. Based on this architecture, we designed a robot welding monitoring system. Considering a large number of different QoS requirements of data flows in IOT, we proposed a kind of load balance routing algorithm based on Glowworm Swarm Optimization-GSOR. In GSOR, the luciferin updating rules of the resident glowworm is mapped to a function with several parameters which reflect the network state and Qos requirements of data flows. The simulation results show that GSOR is able to plan routes according to the given Qos requirements of a data flow and it also outperforms ECMP protocol.

I. Introduction

With the advent of the industrial 4.0 and intelligent manufacturing, Internet-of- Things (IOT) applications have developed rapidly. By 2020, connections of IOT will be more than 2.4 billion sets of equipments [1]. The common IOT applications are based on traditional network architecture, but IP standards cannot deal with the issues of IOT such as the huge number of network objects, huge amounts of information and heterogeneous network structures. At the same time, the distributed traditional network can't truthfully reflect the overall network state. So the effective network strategies are difficult to be formulated. Software Defined Network (SDN) decouples network control and network data, thus it is able to solve the heterogeneity of underlying networks, simplify network configurations and resource management [2]. Furthermore, the centralized structure of SDN enables controller to formulate dynamic network strategies according to the overall network state. Therefore, many problems in traditional networks will be solved through the deployment of SDN on IOT.