Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
IEC 61375-2-3:2015 โ Train Communication Network (TCN) Ethernet Backbone
💡 Key Insight: IEC 61375-2-3 defines the TCN (Train Communication Network) backbone communication profile for modern railway trains. The TCN is a specialised industrial Ethernet architecture that provides deterministic real-time performance, fault tolerance, and interoperability across multiple train consist units in a single communications network.
1. Scope and Network Architecture
IEC 61375-2-3:2015 defines the TCN communication profile for the consist network (backbone) of railway vehicles. As part of the IEC 61375 multi-part standard, this part specifies the Ethernet-based backbone (ETB) that interconnects the individual consist networks (ECN) of each train unit. The standard enables seamless communication between equipment located in different cars of a train, including across coupled train consists in multi-unit operation.
The TCN architecture follows a two-level hierarchy: the Ethernet Train Backbone (ETB) connects the entire train, while each vehicle or consist has its own Ethernet Consist Network (ECN). Gateways between ETB and ECN provide protocol translation, traffic filtering, and network isolation. This hierarchical approach ensures that network problems in one consist do not propagate to the rest of the train.
| Network Segment | Scope | Topology | Data Rate | Max Nodes |
|---|---|---|---|---|
| ETB (Ethernet Train Backbone) | Entire train (multiple consists) | Line/ring, redundant paths | 100 Mbps / 1 Gbps | 256 addresses per train |
| ECN (Ethernet Consist Network) | Single train unit (4-8 vehicles) | Star, line, or ring | 100 Mbps | 63 devices per consist |
| EDN (Ethernet Device Network) | Individual equipment/device level | Star or daisy-chain | 100 Mbps | Deployment-specific |
| Wireless Train Backbone (WTB) | Inter-consist connection (legacy) | Serial bus, redundant | 1.5 Mbps | 32 nodes |
⚠️ Design Consideration: The biggest challenge in TCN backbone design is the train configuration dynamics. Unlike industrial Ethernet networks that have fixed topology, trains can be coupled and uncoupled in the field, changing the network composition and length. The ETB must support automatic discovery, address assignment, and network reconfiguration when consists are joined or separated — all within seconds, without manual intervention, and without interrupting non-critical traffic.
2. Real-Time Communication and Quality of Service
IEC 61375-2-3 defines specific Quality of Service (QoS) mechanisms to ensure deterministic real-time communication over standard Ethernet hardware. Train control functions (traction, braking, door control) require bounded latency and jitter that standard switched Ethernet cannot guarantee without proper configuration.
The standard mandates the use of IEEE 802.1Q VLAN tagging and priority queuing with the following traffic classes:
| Traffic Class | Examples | Priority Level | Max Latency | Message Type |
|---|---|---|---|---|
| Control (Safety-critical) | Emergency brake command, train integrity signal | 7 (Highest) | < 10 ms | Periodic, event-driven |
| Process data | Traction torque demand, door status feedback | 6 | < 50 ms | Periodic (typically 10-100 ms cycle) |
| Supervisory | Diagnostics, fault logging, configuration | 4-5 | < 500 ms | Periodic or on-change |
| Information | Passenger information, CCTV video, audio | 2-3 | < 5 s | Streaming or bulk transfer |
| Best effort | Software updates, file transfers, logs | 0-1 (Lowest) | Not specified | Bulk transfer |
The standard specifies that network switches in the ETB must support strict priority queuing with a minimum of 4 priority queues per port. The use of the IEEE 802.1Q tag’s Priority Code Point (PCP) field enables end-to-end QoS across multi-vendor equipment. Additionally, the ETB must support traffic shaping on gateway ports to prevent message bursts from an ECN from overwhelming the backbone.
✅ Engineering Best Practice: When sizing the ETB switch capacity, always consider the worst-case traffic scenario: train with maximum configured consists (typically 4-8 units), all broadcasting process data at the maximum rate, plus CCTV video streams from passenger cars, plus diagnostic data during fault conditions. A 100 Mbps backbone can handle most operational scenarios, but 1 Gbps is recommended for trains with comprehensive CCTV systems (>16 cameras per consist). Note that the effective throughput after QoS overhead, packet headers, and protocol encapsulation is approximately 60-70% of the nominal line rate.
3. Train Topology Discovery and Network Management
A unique requirement of railway communication networks is the ability to detect and respond to changes in train composition. IEC 61375-2-3 specifies the Train Topology Discovery Protocol (TTDP) that enables automatic detection of consist connections and network reconfiguration. When two trains are coupled, the TTDP completes the following sequence within 2 seconds:
Phase 1 – Physical Detection: The coupler’s electrical or pneumatic connections trigger a physical connection event. Each gateway detects the link-up on its ETB port and initiates TTDP. Phase 2 – Address Assignment: The Train Backbone Manager (TBM) assigns unique IPv6 addresses to all connected consist gateways using stateless address autoconfiguration. Phase 3 – Topology Exchange: Each gateway announces its consist’s capabilities and device inventory to the TBM, which builds a unified train topology map. Phase 4 – Routing Update: The TBM distributes the updated routing table to all gateways, enabling end-to-end communication across the newly formed train. Phase 5 – Application Notification: Higher-level train control functions (TCMS, braking system, passenger information) are notified of the new train configuration and adjust their control strategies accordingly.
🚨 Critical Warning: A poorly designed TTDP implementation creates a security vulnerability. Because TTDP automatically integrates new consists into the network, a malicious or faulty consist could inject false topology information, disrupt routing, or gain unauthorised access to train control functions. IEC 61375-2-3 requires that all TTDP messages be authenticated using the train integrity key, and that gateways implement access control lists (ACLs) to restrict which consists can join the network. These security measures are not optional — they are essential for safe train operation.
4. Redundancy and Fault Tolerance
Railway communication networks must maintain operation in the presence of single-point failures. IEC 61375-2-3 mandates redundant architecture at multiple levels. The ETB must be implemented as a redundant ring topology, where each consist gateway has two ETB ports connected to separate optical fibres that follow different physical paths through the train. Failure of one fibre or one switch must not interrupt backbone communication.
The standard specifies the use of the Media Redundancy Protocol (MRP, per IEC 62439-2) or Parallel Redundancy Protocol (PRP, per IEC 62439-3) for network-level redundancy. MRP provides ring recovery times below 20 ms (typically 3-10 ms for 10-switch rings), while PRP offers zero-switchover time by duplicating all packets over two independent networks. For safety-critical control data, PRP is the preferred approach despite its higher infrastructure cost (two complete networks).
💡 Practical Recommendation: For new train designs, use PRP (IEC 62439-3 Clause 4) for the ETB rather than MRP. The cost difference is primarily in additional Ethernet ports and cabling (approximately 15-20% more per node), but the benefit of zero-loss switchover during network faults is critical for train control functions. MRP’s 3-20 ms recovery time, while acceptable for most process data, may cause missed messages in safety-critical chains where SIL4 response times are in the 10-50 ms range.
5. Frequently Asked Questions
Q1: What is the relationship between IEC 61375-2-3 and IEC 61375-1?
A: IEC 61375-1 is the foundational part of the TCN standard, defining the overall architecture, functional requirements, and communication principles. IEC 61375-2-3 is a specific profile within Part 2 (the Ethernet-based train backbone), providing the detailed protocol specification, QoS configuration, and topology management procedures for the ETB. Both are required for a complete TCN implementation.
Q2: Can standard industrial Ethernet switches be used in the ETB?
A: No — standard industrial Ethernet switches lack several features required by IEC 61375-2-3, including: TTDP support, multiple MRP instances, train-specific QoS mapping, and the extended temperature and vibration rating for rail applications. Switches must be specifically qualified as TCN-compliant. However, standard Ethernet physical layers (100BASE-TX, 100BASE-FX, 1000BASE-LX) are used, so standard cabling and connector practices apply.
Q3: How is timing synchronisation handled across the TCN?
A: The standard uses IEEE 1588v2 (Precision Time Protocol) for clock synchronisation across the train. The TBM acts as the grandmaster clock, and all gateways synchronise to it via the ETB. Accuracy requirements are typically below 1 microsecond for time-stamping of event data. Each ECN can operate its own IEEE 1588v2 domain synchronised to the ETB time.
Q4: Does IEC 61375-2-3 support legacy TCN (WTB/MVB) integration?
A: Yes. The standard provides for gateway-based integration between ETB and legacy TCN networks (WTB for inter-consist, MVB for intra-consist). The consist gateway translates between ETB protocols and WTB/MVB telegrams. However, the gateway introduces additional latency (typically 1-10 ms), which must be accounted for in end-to-end timing analysis for control functions spanning both network types.
