Given their nature, many control applications that arise from the integration of Operation Technologies (OT) and Information Technologies (IT) are built on top of highly reliable real-time (RT) Distributed Control Systems (DCSs). Since a DCS is made up of several computing nodes that exchange information through a communication subsystem, to achieve high reliability it is necessary that both this subsystem and the service from the nodes are very reliable. To provide RT highly reliable communications while benefiting from Ethernet's advantages, Industry and Academia are pushing the Time-Sensitive Networking Ethernet standards (TSN). On the other hand, one of the most used strategies to ensure a highly reliable service from the nodes is to use fault tolerance in the form of active replication. Our general goal is to develop a complete fault-tolerant architecture (addressing faults both in the communication subsystem and in the nodes) for highly reliable real-time DCSs based on TSN. In this paper we show our ongoing work towards an active replication schema for the nodes of this architecture.

Modern electrical Substation Automation Systems (SAS) are designed following the guidelines defined in the IEC 61850 standard. This standard specifies the necessary information models and communication services for SAS in such a way that they are independent of the implementation. This allows system designers to choose the specific communication technology that best fits their needs. Time-Sensitive Networking (TSN) is arising as one of the most appealing technologies for this purpose. However, it is necessary to map the communication services to TSN, ensuring that the real-time and fault-tolerance requirements of the messages are met. In particular, efficiently mapping the messages of the Generic Object Oriented Substation Events (GOOSE) service is challenging. This is because it exhibits a transmission pattern that does not align with the types of traffic defined in TSN. In this paper we analyze this transmission pattern, identify and characterize its subpatterns, propose the most suitable mapping for each of them and discuss the efficiency gain with respect the typical approaches used to do this mapping.

New emerging Distributed Control Systems (DCSs), like Substation Automation Systems (SASs) of Smart Power Grids, raise new requirements on their underlying control networks. To meet these new requirements, both Industry and Academia are promoting the Time-Sensitive Networking (TSN) Ethernet standards. In particular, TSN includes mechanisms to exchange information simultaneously through several paths of practically any spatially redundant network topology. This topological flexibility can offer a better balance between fault tol-erance (FT) and redundancy cost (extra number of components) than classical Industrial Ethernets. However, the mentioned TSN mechanisms may also increase the cost in terms of extra latency and jitter, which could jeopardize real-time communications. In this paper we show our ongoing work to experimentally assess this extra latency and jitter and, thus, characterize the benefits of TSN in terms of balance between FT and cost.

Electrical substations are vital for the power grid, and Substation Automation Systems (SASs) have been employed to enhance substation functionality and safety. As the energy landscape evolves, substations face new challenges such as accommodating an increasing number of prosumers. Thus, SASs require a reliable substation communication network (SCN) capable of supporting real-time control and diverse applications. While Ethernet-based SCN technologies have emerged, they often fall short in meeting all requirements, including TCP/IP support, cost-effective fault tolerance, and managing traffic with different real-time demands. Time-Sensitive Networking (TSN) standards have shown promise in addressing these limitations by providing novel mechanisms. In this paper we compare TSN with the Parallel Redundancy Protocol (PRP) demonstrating that TSN offers better functionality and efficiency. In the direction of designing a comprehensive TSN-based architecture for SASs’ Distributed Control Systems (DCSs) we start here by proposing a roadmap for the fault tolerance aspects.

Distributed Control Systems (DCSs) for emerging industrial control applications impose new communication requirements that cannot be satisfied by current Industrial Ethernet protocols. As a result, industry is pushing the Time-Sensitive Networking (TSN) standards as the de-facto Ethernet-based linklayer to fulfill these requirements. Adequate roadmaps are needed to support a smooth transition from Industrial-Ethernet-based legacy systems to TSN-based ones. In this context some works propose mechanisms to migrate, i.e. map, route and schedule, legacy traffic to TSN. However none of them considers traffic including streams with spatial redundancy requirements and, thus, they cannot be used to migrate legacy highly-reliable DCSs. The present work extends a previous toolchain to migrate, for the first time, legacy critical traffic that includes spatially redundant streams. Particularly, since redundancy is costly, this work proposes and compares two routing methods that consider one redundant stream per traffic.