PRP

What is PRP?

PRP (Parallel Redundancy Protocol) is an Ethernet redundancy protocol defined in IEC 62439-3 in which every device sends each frame simultaneously over two completely separate networks, so that the failure of one network costs not a single frame and the recovery time is zero. The receiver uses the first copy that arrives and discards the duplicate. PRP is used wherever even an interruption of a few milliseconds is unacceptable, such as digital substations based on IEC 61850, power supply, railways and critical process automation.

🧠 How does PRP work?

A PRP network consists of two parallel, independent networks: LAN A and LAN B. There is no connection at all between them, so a fault in one cannot affect the other. The redundancy does not live in the switches but in the end devices:

1. Dual attachment - a PRP device has two network ports, one to LAN A and one to LAN B
2. Dual transmission - the PRP layer in the device sends every frame through both ports at the same time
3. Tagging - a Redundancy Control Trailer (RCT) with a sequence number and the network identifier is appended to each frame
4. Accept the first, discard the second - the receiver recognises the pair by source address and sequence number, passes the first copy to the application and drops the second
5. No switchover - if LAN A fails, frames simply keep arriving via LAN B; there is no failover and therefore no recovery time

Because both networks are always active, the application does not notice a cable break or a failed switch. That makes PRP suitable for traffic that cannot tolerate packet loss, such as GOOSE messages between protection relays, and for applications with strict real-time requirements.

🔧 Which devices and frame elements does PRP use?

The trailer sits at the end of the frame. Devices without PRP treat those six bytes as padding and ignore them, so ordinary devices and PRP devices can share the same network. Switches and older devices must, however, accept frames of up to 1528 bytes, because the RCT makes a maximum-size frame slightly longer.

Because every device receives supervision frames, management software sees immediately when a path fails. That matters: a fault in LAN A is invisible to the application, but it must be repaired quickly, otherwise the installation keeps running without redundancy and nobody notices.

🕰️ How has PRP evolved?

PRP and its ring-shaped sibling HSR were developed by suppliers to the power industry, including Hirschmann, ABB, Siemens, Alstom and RuggedCom (now part of Siemens).

PRP-0 and PRP-1 use different frame formats. New equipment uses PRP-1; when you extend an older installation, check which version the existing devices speak.

🔄 What is the difference between PRP, HSR, MRP and RSTP?

PRP and HSR send everything twice and therefore never switch over. MRP, RSTP and DLR block a path and only activate it when a fault occurs, so frames are lost during recovery.

PRP is flexible: each of the two networks may have its own topology and may itself use RSTP or MRP internally. HSR saves switches, but the ring carries all traffic twice. Larger installations combine both through RedBoxes. A broader overview is available under Ring Redundancy.

🏭 Where is PRP used in practice?

• Digital substations - IEC 61850 names PRP and HSR as the redundancy protocols for the station and process bus; IEDs exchange protection messages and measurements over them
• Power and grid operation - control of high-voltage substations, where communication with SCADA and protection must never falter
• Process industry - control networks around DCS and PLC systems in chemicals and oil and gas
• Synchronised drives - Drives in printing presses and power inverters that cannot tolerate a missed cycle
• Rail and infrastructure - signalling systems and tunnel control

PRP operates at layer 2 and is transparent to the protocols above it. Anything running over Industrial Ethernet, from OPC UA to IEC 61850 GOOSE, therefore benefits without modification. In the Purdue Model you mainly find PRP at levels 0 to 2: field, control and supervisory networks.

For time-critical applications the clock must be redundant too. IEC 62439-3 therefore contains, in Annex C, a profile for the Precision Time Protocol over PRP and HSR. The power-industry variant was adopted as IEC/IEEE 61850-9-3 and achieves an accuracy of about 1 microsecond.

🧭 How do you design a PRP network?

1. Decide what needs to be redundant - only traffic that cannot tolerate any interruption justifies a duplicate network; the rest can use MRP or RSTP
2. Build two truly independent networks - separate switches, separate power supplies and preferably separate cable routes, so that one fire or excavation incident cannot hit both
3. Never connect LAN A and LAN B - a link between the networks causes loops and duplicate frames and defeats the redundancy
4. Connect single-port devices through a RedBox - engineering laptops, cameras or older relays
5. Check frame size and version - switches must pass frames of up to 1528 bytes; do not mix PRP-0 and PRP-1 unknowingly
6. Monitor both networks - use the supervision frames and the devices’ error counters; a failed path must raise an alert
7. Test at commissioning - pull a cable from LAN A and then from LAN B, and confirm that the application notices nothing while the monitoring does raise an alarm

Keep in mind that PRP does not protect against failure of the end device itself. That requires redundant controllers or high-availability solutions. Every additional network component can also fail; because faults stay invisible to the application, regular checks and maintenance become more important, not less.

🔐 How do you secure a PRP network?

PRP increases availability, but not security. The protocol has no authentication or encryption, and an attacker with access to one of the networks can forge supervision frames or traffic.

• Secure both networks equally - with firewalls, ACLs and VLANs; a weaker LAN B is an open back door
• Monitor both channels - send logs and alerts to a SIEM or OT SOC and run network monitoring on both networks; Wireshark can dissect PRP supervision frames
• Restrict access - with RBAC for managing switches and RedBoxes, plus physical protection of cabinets and unused ports
• Apply segmentation - following the zones and conduits of IEC 62443, with both PRP networks in the same zone
• Keep configurations identical - apply and document changes on LAN A and LAN B together

Grid operators and many other energy companies in the Netherlands fall under the Cybersecurity Act (Cyberbeveiligingswet), in force since 15 August 2026. A redundant network helps meet continuity requirements, but it does not replace a risk assessment or security measures. Only a well-managed redundant network genuinely supports safety and reliability.

❓ Frequently asked questions

What does PRP stand for?

PRP stands for Parallel Redundancy Protocol, an Ethernet redundancy protocol defined in IEC 62439-3. PRP sends every frame simultaneously over two separate networks, so the failure of one network causes no interruption.

Does PRP really have zero recovery time?

Yes, PRP has zero recovery time because it never needs to switch over. Both networks carry all traffic permanently, and the receiver simply uses the copy that arrives through the working network.

What is a RedBox in PRP?

A RedBox (Redundancy Box) is a device that connects ordinary single-port devices to both PRP networks. The RedBox duplicates their frames and removes duplicates, so devices without PRP support also become redundantly reachable.

Can LAN A and LAN B be connected in a PRP network?

No, the two PRP networks must not be connected anywhere. A link causes loops and duplicate traffic and removes their independence, so a single fault could still affect both paths.

What is the difference between PRP and HSR?

PRP uses two separate networks with standard switches, while HSR sends frames in both directions around a single ring of HSR devices. Both come from IEC 62439-3 and have zero recovery time; PRP scales better, and HSR is cheaper because it needs no second network.

Does PRP work with every industrial protocol?

Yes, PRP operates at layer 2 of the OSI model and is therefore transparent to higher-level protocols. IEC 61850, OPC UA, Modbus TCP and other Ethernet-based protocols benefit from PRP without the application needing to know about it.

📌 In summary

PRP is the Ethernet redundancy protocol from IEC 62439-3 that sends every frame over two separate networks and therefore keeps working, with zero recovery time and no packet loss, when one network fails. The price is a fully duplicated network; design both networks to be truly independent, monitor both and secure them equally, because PRP increases availability but not security.