A Confguration Agent for Dynamic Time-Critical Networks

Traditionally, the configuration processes of real-time networks are costly both in terms of time and engineering efiorts, and typically require the system to be shutdown during the reconfiguration phase, thus resulting in potentially costly down time. Whereas this characteristic has not been a major issue in the past due to the static nature of Operational Technologies (OT) networks, the online network reconfiguration is now becoming a challenging demand for the emerging contexts of the Real- Time Internet of Things (RT-IoT), Industry 4.0 or smart cities. These new networks need to accommodate a much higher number of participants which makes changes in the network, in the functionality or for maintenance purposes, much more frequent. Shutting down the system to adapt the configuration to any new change would be then, not feasible due to the real-time characteristics of such systems. Thus the network must be capable to adapt to changes while still being able to provide deterministic guarantees for the most critical trafic. The solution that we propose is to use some advanced characteristics of Information Technologies (IT) for the OT world by considering the configuration as a continuous service for the network. In order to realize this vision we introduce the Configuration Agent: an autonomous entity that learns the characteristics of the network through continuous monitoring, with the goal of facilitating the configuration and re-configuration of heterogeneous trafic profiles. With an SDN-like architecture that has a global view of the network, the Configuration Agent is applicable in wider contexts, such as security, safety or wireless.