Bjarne Johansson

Revisiting Spatial Redundancy in Industrial Controller Architectures: A Network-Centric Perspective

Student   Bjarne Johansson
Advisors   Thomas Nolte
Alessandro V. Papadopoulos
Faculty Reviewer   Thilo Sauter, University for Continuing Education Krems, Austria
Grading Committee   Ahlem Mifdaoui University of Toulouse/ ISAE-Supaéro, Toulouse, France
Luis Almeida, University of Porto, Portugal
Paul Pop, DTU, Copenhagen, Denmark_
Kristina Lundqvist, Mälardalen University, Sweden (reserve)
Defence   Mälardalen University, Västerås, Sweden
Room Kappa and Teams/Zoom meeting (Link will be made public)
November 6th, 2025 13:15
Abstract   Automation solutions are omnipresent in modern society as a part of the infrastructure that provides utility services such as water and power. At the core of these systems is the controller, a specialized computer designed to operate in harsh environments where unplanned downtime can be costly. High-quality hardware, software, and spatial redundancy (i.e., hardware multiplication) are commonly employed to mitigate disruptions.
Industrial control systems are evolving into more interconnected and interoperable architectures, marking a shift toward network-centric designs where the network, rather than the controller, becomes the central part of the system. Concepts traditionally associated with information technology, such as edge and cloud computing, containerization, and orchestrators, are entering the operational technology domain. New standards, such as OPC UA, with its information model and communication protocols, are gaining traction to facilitate interoperability.
This evolution presents redundancy challenges, such as adapting failure detection and state transfer mechanisms needed by standby redundancy to a network context, and opportunities, such as utilizing systems previously confined to the information technology domain. This shift toward a network-centric control system architecture is the overarching motivation for this thesis’s revisit of spatial redundancy.
Specifically, this thesis investigates orchestrator-aided failure recovery as a complement to traditional redundancy. It also proposes a failure detection mechanism that maintains consistent control during network partitioning between redundant controllers. The thesis also examines the behavior of OPC UA PubSub in a standby redundancy context. It introduces a method for processing priority based on information embedded in incoming network frames. Additionally, the thesis proposes an architecture that enables the distribution of redundancy-related state data. It also investigates checkpointing solutions and communication protocols to identify a suitable mechanism for transferring state data between redundant controllers.
Rules and Guidelines   The PhD procedure summary
Guidelines for Third-Cycle Studies at MDU
Thesis   Thesis
Included Papers   Paper A: Kubernetes Orchestration of High Availability Distributed Control Systems .
Paper B: Consistency Before Availability: Network Reference Point based Failure Detection for Controller Redundancy .
Paper C: OPC UA PubSub and Industrial Controller Redundancy .
Paper D: Priority Based Ethernet Handling in Real-Time End System with Ethernet Controller Filtering .
Paper E: Partible State Replication for Industrial Controller Redundancy .
Paper F: Checkpointing and State Transfer for Industrial Controller Redundancy.
Publications   Complete list of publications

Back to Research

Last modified: 2025-09-01 11:16:54 +0200