A key component of realizing Industry 4.0 is achieving deterministic networking performance across all levels of performance. This talk examines how network stack choices impact latency SLAs, performance, and scalability. We compare AF_XDP and DPDK on Linux systems, showcasing how application developers can leverage their unique properties with the underlying hardware platforms. A design guide is provided to help select the right network stack for future SLA-driven deterministic deployments.

Fault-tolerant time synchronization is crucial within converged 6G-TSN networks, particularly for future industrial automation scenarios that demand deterministic communication. We present a comparison of hot standby (IEEE 802.1ASdm) with BTCA to enable fault tolerant time synchronization for converged networks. We also introduce our OMNeT++ simulation framework used for the analysis and explore various industrial automation use cases that stand to benefit from the adoption of a hot standby GM.

Inspired by the success of the Avnu Express Test System (AETS) we have embarked on a quest to push the limits of streaming Milan/AVB audio on a standard PC running the Linux operating system. We share our experience and lessons learned focusing on requirement for hardware and application requirements, configuration options and engineering challenges. We conclude with performance test results and provide ideas for possible next steps.

The integration of 5G and TSN introduces significant packet delay and delay variation making it challenging to compute 802.1Qbv schedules with latency and reliability guarantees. We provide an overview of strategies employing statistical knowledge of 5G channel conditions to allocate suitable 5G packet delay budgets (PDBs). We discuss the trade-off between choosing 5G PDBs tightly, requiring online reconfiguration of schedules, and conservative 5G PDBs, resulting in poor latency guarantees.

IEEE 802.1AS-2024 defines a method for clock drift mitigation, specifically designed for systems with many hops between Grandmaster and end stations. In this talk, we give insights into the reasoning behind this feature and present real-world results using the Drift_Tracking TLV and its handling in PTP relays and end stations. Different scenarios for drift in Grandmaster and PTP relays complement the simulations done at IEEE in the context of the Industrial Automation profile IEC/IEEE 60802.

The demo showcases a wireless time-sensitive networking solution using the openwifi platform. It highlights W-TSN features like over-the-air time synchronization, time-aware shaping, deterministic handover, and a network management controller. The demo involves a rover navigating an area covered by two W-TSN enabled APs, demonstrating deterministic handover utilizing a handover controller and UWB-based localization. A visual app will show seamless transitions with the proposed handover solution.

In order to take full advantage of virtualization, time-sensitive applications must be integrated with 10G+ network interface cards and related platforms. In this presentation, we showcase practical results using latest Intel hardware running on Linux. PROFINET communications are benchmarked using the Linux Real Time Communications Testbench. We highlight the features employed to optimize transmission jitter in several orders of magnitude, and some of the challenges ahead.

Modern vehicles need more bandwidth for data processing and communication, especially for autonomous driving functionalities. Our research aims to create a Time-Sensitive Networking (TSN) environment for zonal architecture car communication, integrating 10Gbps TSN Endpoints and Switches. By combining radar sensor technology with TSN and Precision Time Protocol, our findings significantly enhance vehicle capabilities, promoting safer and more efficient intelligent transportation systems.

This manuscript elaborates on the OPC UA FX adoption of IEEE TSN as a transport layer, reflecting on UAFX integration of the CNC/CUC concept and network configuration. It particularly highlights the UAFX Connection Manager, configuration workflows, and OPC UA information modelling used in devices for abstracting the existing traffic management configuration with QoS concepts that can abstract and incorporate not just existing, but also upcoming mechaninsms.

During the transition to IEEE 802.1AS-2020, many TSN setups will mix AS-2020 and legacy AS-2011 devices. AS-2020 Grandmasters are still limited, but also AS-2011 switches or endnodes can't always be upgraded. A key change in AS-2020 is the peer delay mechanism. To ensure compatibility, AS-2020 offers configurable mechanisms. Our talk explores maximizing interoperability without explicit port configs, presenting real-world-tested strategies for robust time sync in industrial TSN environments.

With IT/OT convergence in industrial networks, Time-Sensitive Networking (TSN) ensures reliable operations, requiring configuration for routers, switches, and TSN-Endpoints. Our research explores TSN configuration for IoT devices with cost-effective hardware. NETCONF manages robust devices, while CORECONF targets constrained ones. We propose integrating NETCONF with CORECONF via augmented YANG models, improving network reliability and simplifying TSN deployment for constrained devices.

Adding time-sensitive network sections to existing installations is a major use case across industries. It leads so significant challenges regarding synchronization. This study explores wired and wireless solutions, including 5G, to interconnect TSN sections while maintaining precise timing. Various approaches are evaluated for feasibility and performance using accurate measurement setups, demonstrating effective methods to minimize sync errors and extend TSN beyond dedicated infrastructures.

This presentation showcases proven migration strategies for TSN using PROFINET infrastructures. By combining a Linux-based PROFINET controller with a TSN switch, we enable parallel operation of multiple PROFINET domains without modifying existing end devices. This creates a converged network platform for real-time and best-effort communication. We demonstrate how TSN scales industrial throughput via PROFINET while keeping the path open for future device generations with enhanced TSN features.

It becomes evident that TSN adoption will happen over a period of time where new technology will coexist with legacy Ethernets and fieldbuses.
We would like to challenge the existing testbeds, plugfests, and IA vendors with accommodating the missing items, including support for most used industrial RTOSes, generic orchestration SW, and HW platforms, suitable to run most of the Industrial Automation use cases.

Dividing a TSN network into multiple domains is attractive for many reasons: (1) autonomous management; (2) increased scalability (3) separation of different technologies (wireline/-less); (4) fault tolerance and service protection (isolation, redundancy). Besides motivating multi-domain systems, we introduce and discuss aspects required to realize multi-domain TSN systems, such as an architecture, control/management aspects, inter-domain notifications, aggregation, and service protection.

Broader application ranges in converged networks can increase the presence of heterogeneous traffic with mixed criticality, which raises the needs for effective mechanisms protecting against disruptive network traffic from misbehaving devices. The TSN toolbox provides multiple policing mechanisms intended for this purpose in different systems (e.g., automotive in-vehicle or aviation). This presentation provides insights into needs, function and practical use of the TSN policing mechanisms, and shows relationships with the TSN shaping & scheduling mechanisms.

This presentation summarizes some recent TSN developments in China and highlights critical gaps: the shortage of TSN-capable end-stations and the lack of proven solution practices. We introduce the TSN Autolearning mechanism which enables brownfield deployment through switch-based implementation and analyze its limitations. We then discuss the need for TSN to evolve into a generic, user-friendly ICT infrastructure, supported by multi-vendor ecosystems and open communities. Drawing on Ethernet's evolutionary trajectory and two case studies, we illustrate why a mindset shift is necessary and offer suggestions.