Abstract
Industrial Ethernet rings are designed to eliminate single points of failure.
But simply connecting switches into a loop will immediately create broadcast storms unless a protection mechanism exists.
ERPS (Ethernet Ring Protection Switching), defined in ITU-T G.8032, provides a deterministic, fast, and vendor-independent method to recover from link or switch failures—typically within 50 ms.
This article explains how ERPS works, why it is widely adopted in industrial automation, and how it differs from other redundancy protocols.
Why Do Industrial Networks Use Rings?
Unlike office networks, industrial facilities often require:
-
Long-distance fiber connections
-
High availability
-
Simple cabling
-
No single point of failure
Typical applications include:
-
Airports
-
Water Treatment Plants
-
Power Utilities
-
Oil & Gas
-
Rail Transit
-
Manufacturing
Instead of star topology, engineers frequently build a fiber ring.
Unfortunately… Ethernet does not allow loops.
Without protection:
-
Broadcast Storms
-
MAC Address Flapping
-
Multiple Frame Copies
-
Network Collapse
A ring protocol is therefore required.
What is ERPS?
ERPS stands for
Ethernet Ring Protection Switching
It is standardized in
ITU-T G.8032
Unlike proprietary ring protocols, ERPS allows switches from different vendors to participate in the same ring (provided they implement the standard correctly).
Its goal is simple:
-
Keep one path intentionally blocked.
-
Immediately unblock it when a failure occurs.
Basic Working Principle
Imagine the five-switch fiber ring above.
Normally, one link is blocked.
Traffic only travels in one logical loop-free path.
The blocked link is called the
RPL (Ring Protection Link)
The Ring Protection Link (RPL)
Every ERPS ring contains one special node.
It is called the
RPL Owner
The RPL Owner intentionally blocks one ring port.
This creates a loop-free topology.
Under normal operation:
-
Traffic →A → B → C → D → E
-
Blocked E ✕ A
The blocked port carries no user traffic.
What Happens When a Fiber Breaks?
Suppose the fiber between B and C is cut.
-
Before: A → B → C → D → E
-
After failure: A → B X C → D → E
The ring becomes disconnected.
ERPS immediately detects:
-
Loss of Signal
-
Link Down
The RPL Owner receives the event.
It instantly opens the previously blocked port.
-
Now traffic flows: A ← E ← D ← C
The ring is restored automatically.
Recovery is typically
<50 ms
How Does ERPS Detect Failures?
Each ERPS node periodically exchanges
R-APS (Ring Automatic Protection Switching)
control messages.
These messages report:
-
Link status
-
Ring health
-
Protection state
If an R-APS message changes unexpectedly, nodes update their forwarding state.
Unlike STP, the entire network does not need to recalculate a topology.
Only the ring changes state.
This is why ERPS is much faster.
ERPS Recovery Process
Normal
Ring Closed
↓
One Port Blocked
Fiber Failure
Link Down Detected
↓
R-APS Sent
↓
Blocked Port Opens
↓
Traffic Restored
↓
Repair Completed
↓
Wait-To-Restore Timer
↓
Original Block Restored
Why is ERPS Faster than STP?
Typical Industrial Applications
ERPS is commonly deployed in:
-
Airport communication networks
-
ITS roadside cabinets
-
Metro Ethernet
-
Utility substations
-
Manufacturing plants
-
SCADA fiber networks
-
Pipeline monitoring
-
Tunnel communication systems
Engineering Considerations
When deploying ERPS:
Only one RPL Owner should exist per ring.
Ring ports should not simultaneously participate in STP on the same ring, as the mechanisms can conflict.
Use identical VLAN membership on all ring ports.
Configure timers consistently across all switches in the ring.
Test recovery by disconnecting a fiber and verifying that traffic is restored within the required application tolerance.