Exactly one MRM per ring
Two MRMs cause continuous ring oscillation. Zero MRMs means no protection. Configure exactly one.
11.1 MRP Overview
The previous chapter covered TSN, the future of deterministic Ethernet. Today’s industrial networks rely on a proven, simpler mechanism for redundancy: MRP. This chapter explains how MRP works at the protocol level, from normal operation through failure and recovery.
Why MRP Exists
Section titled “Why MRP Exists”STP and RSTP were designed for arbitrary mesh topologies. On a ring, RSTP converges in 1 to 2 seconds. That is too slow for industrial automation.
| Protocol | Topology | Convergence | Suitable for OT? |
|---|---|---|---|
| STP (802.1D) | Mesh | 30 to 50 s | No |
| RSTP (802.1w) | Mesh / Ring | 1 to 2 s | Marginal |
| MRP (IEC 62439-2) | Ring | < 200 ms | Yes |
A 1-second outage causes PLCs to fault. A 200 ms outage is within the watchdog tolerance of most industrial controllers. MRP (Media Redundancy Protocol), defined in IEC 62439-2, is purpose-built for ring topologies and achieves sub-200 ms recovery.
Ring Structure
Section titled “Ring Structure”An MRP ring is a closed loop of switches. Exactly one switch is the MRM (Media Redundancy Manager). All others are MRC (Media Redundancy Client) nodes.
Under normal operation, the MRM blocks its secondary port. The ring is logically open at that point. Traffic flows in one direction around the ring. The MRM sends MRP_Test frames every 10 to 20 ms to verify ring integrity. If the MRM receives its own test frames back, the ring is intact.
All switches in the same ring share a Domain UUID. All ring ports on every switch belong to the MRP VLAN. These two settings identify the ring and carry its control frames.
Normal Operation
Section titled “Normal Operation”The MRM sends test frames on both ports. The secondary port is blocked, so test frames arriving there are discarded. The MRM acts only if test frames stop arriving on the primary port.
Failure Detection and Recovery
Section titled “Failure Detection and Recovery”When a link fails, the adjacent MRC sends an MRP_LinkDown frame to the MRM immediately (less than 1 ms). The MRM unblocks its secondary port and sends MRP_TopologyChange to all ring members, triggering a MAC table flush.
When the failed link is repaired, the MRC sends MRP_LinkUp. The MRM blocks its secondary port again and sends another MRP_TopologyChange. Normal operation resumes.
MRP and VLANs
Section titled “MRP and VLANs”MRP control frames travel in the MRP VLAN. Every ring port on every switch belongs to this VLAN. A ring port not in the MRP VLAN silently breaks the ring. The MRM reports the ring as open even though all cables are connected.
Configuring MRP on Hirschmann HiOS
Section titled “Configuring MRP on Hirschmann HiOS”Navigate to Switching → L2-Redundancy → MRP and configure:
| Parameter | Value |
|---|---|
| Domain UUID | Same on all ring switches |
| MRP VLAN | Same on all ring switches |
| Role | Manager (one switch only) / Client (all others) |
| Ring Port 1 | First ring port |
| Ring Port 2 | Second ring port |
After configuration, verify: the MRM shows Ring state: Closed.
Key Takeaways
Section titled “Key Takeaways”All ring ports need the MRP VLAN
A ring port not in the MRP VLAN silently breaks the ring.
Same Domain UUID on all ring switches
The UUID identifies the ring. Mismatched UUIDs mean switches ignore each other’s MRP frames.
What Comes Next
Section titled “What Comes Next”This chapter explained how MRP works at the protocol level. The next chapter goes deeper: the wire format of every MRP frame type, the MRM state machine, and Python tools for parsing MRP frames and monitoring ring health in real time.
References
Section titled “References”- IEC 62439-2:2016 — Media Redundancy Protocol (MRP)
- Hirschmann. (2023). User Manual: HiOS MRP Configuration. Belden/Hirschmann.
- PROFIBUS & PROFINET International. (2020). MRP Application Note.