An Enhanced Reconfiguration for Deterministic Transmission in Time-Triggered Networks
Zonghui Li,Hai Wan,Zaiyu Pang,Qiubo Chen,Yangdong Deng,Xibin Zhao,Yue Gao,Xiaoyu Song,Ming Gu
DOI: https://doi.org/10.1109/TNET.2019.2911272
2019-06-01
IEEE/ACM Transactions on Networking
Abstract:The emerging momentum of digital transformation of industry, i.e. Industry 4.0, poses strong demands for integrating industrial control networks, and Ethernet to enable the real-time Internet of Things RT-IoT. Time-triggered TT networks provide a cost-efficient integrated solution while RT-IoT arouses the reconfiguration challenges: the network has to be flexible enough to adapt to changes and yet provides deterministic transmission persistently during network reconfiguration. Software defined network benefits the flexible industrial control by configuring the rules handling frames. However, previous reconfiguration mechanisms are mostly oriented to the context of data centers and wide area networks and thus do not consider the deterministic transmission in TT networks. This paper focuses on the reconfiguration i.e., updates for the deterministic transmission. To minimize the overhead during updates, namely the minimum number of loss frames and the minimum duration time of updates, we first establish an update theory based on the dependence relationship derived by the conflicts during updates. In addition then the reconfiguration problem is modeled with the dependence graph built by the relationship. On such a basis, we present a reconfiguration mechanism and its implementation to solve the problem. Finally, we evaluate the proposed reconfiguration mechanism in two real industrial network topologies. The experimental results demonstrate that compared with previous methods, our mechanism significantly reduces the number of loss frames and achieves zero loss in almost all cases.