Powerlink

What is Powerlink?

Powerlink, in full Ethernet POWERLINK, is an open industrial Ethernet protocol that makes standard Ethernet (IEEE 802.3) deterministic by letting a single coordinator, the Managing Node, decide in every cycle which device may transmit and when. As a result there are no collisions or queues, and the network reaches cycle times below 200 µs with jitter of less than 1 µs. Powerlink is mainly used for real-time motion control, fast I/O and machine control, and is standardised internationally in IEC 61158 and IEC 61784-2.

🧠 Where does Powerlink come from and who manages it?

The Austrian automation company B&R introduced Powerlink in 2001. In June 2003 the independent Ethernet POWERLINK Standardization Group (EPSG) was founded to develop the protocol further as an open standard, in cooperation with CAN in Automation (CiA) and the IEC. The EPSG had working groups for safety, technology, certification and end users.

Two developments changed how it is governed:

• 2017 - ABB acquired B&R, which has since formed the core of ABB’s machine automation business.
• 31 March 2023 - the EPSG was dissolved. B&R is its formal successor and has carried full responsibility for development, support and certification since April 2023. The specifications remain publicly available, and existing products and certificates remained valid.

Powerlink therefore remains an open protocol: anyone may implement it without licence fees, and openPOWERLINK is an open-source stack under a BSD licence that can run as master or slave on Linux, Windows and VxWorks, among others.

🔧 How does the Powerlink cycle work?

Ordinary Ethernet devices transmit whenever they like. On a shared medium this causes collisions (CSMA/CD), and on a switch it causes queuing; both are unpredictable. Powerlink solves this with a mix of polling and time slicing: only the Managing Node (MN), usually the PLC or motion controller, grants the right to transmit. All other devices are Controlled Nodes (CN) and transmit only when they are told to.

Each cycle consists of fixed frames:

The cycle therefore has an isochronous phase for process data and an asynchronous phase for everything else. Because every PRes is a multicast, a producer-consumer model emerges: one servo drive can receive the position of another axis directly, without a detour via the controller. That is useful for electronic gearing and cam profiles.

Two optimisations make the network more efficient:

• Multiplexing - less critical nodes share a time slot and are polled every few cycles, so the basic cycle stays short for the fast axes
• Poll Response Chaining - CNs reply one after another to a single request, reducing the number of frames in large installations and robotics

The application layer is derived from CANopen: an object dictionary with a 16-bit index and up to 256 sub-indices, plus the familiar CANopen device profiles such as CiA 402 for drives. Every device is described in an XDD file according to ISO 15745-4.

📊 What are Powerlink’s typical values and properties?

Powerlink works on layers 2 and 7 of the OSI model and uses ordinary Ethernet chips. Within the real-time domain, however, the specification recommends repeating hubs over switches, because a hub adds less delay and jitter; the MN already prevents collisions through its polling schedule. Topologies can be chosen freely: line (daisy chain through devices with a built-in hub), star, tree or ring. See also Topology.

For availability Powerlink offers hot plugging (a new or replaced node announces itself to the MN in the asynchronous phase), ring redundancy (one extra cable closes the line into a ring if the MN has two redundancy-capable ports) and a redundant MN in hot standby that takes over immediately on failure. More background is in Redundancy.

🔄 How does Powerlink compare with other real-time protocols?

Powerlink and SERCOS III are similar because both divide the medium into fixed time slots. The main difference from EtherCAT is that Powerlink needs no special slave hardware, while IP relies on ordinary switched Ethernet with time synchronisation. Market analyses give Powerlink a small share next to PROFINET, EtherNet/IP and EtherCAT, with a strong position on machines using B&R controllers. For new converged networks many machine builders also look at TSN and OPC UA FX. An overview of older technology is in Fieldbuses.

🏭 Where is Powerlink used?

For safety functions there is openSAFETY, formerly POWERLINK Safety. This safety protocol wraps safety data in ordinary messages (black channel), is TÜV-certified for SIL 3 according to IEC 61508 and is standardised as IEC 61784-3-13. It is network-independent and therefore an alternative to CIP Safety. See also Functional Safety and Safety PLC.

🔐 How do you secure a Powerlink network?

Powerlink runs on layer 2 Ethernet and has no authentication or encryption of its own. Any device that sends frames into the real-time domain can disrupt the polling schedule. Security therefore has to come from the architecture:

1. Isolate the real-time domain - keep office and IT traffic physically out of the Powerlink segment; a foreign transmitter breaks the timing.
2. Connect via a gateway or controller - let only the MN or a gateway with two network ports connect to the machine network, and place a Firewall between machine and plant network according to the zones and conduits model of IEC 62443.
3. Segment the networks above - apply network segmentation, for example with VLANs, so an infected workstation cannot reach machines.
4. Monitor passively - use IDS sensors and feed alerts into a SIEM; Wireshark can decode Powerlink frames for diagnostics.
5. Control access - restrict who can change configurations from an engineering station.

Regulation pushes in the same direction. Machine builders face the Cyber Resilience Act for products with digital elements and the Machinery Regulation, which from 20 January 2027 makes protection against corruption of control systems a safety requirement. Their customers in essential and important sectors fall under the Dutch Cybersecurity Act (Cyberbeveiligingswet), the national implementation of NIS2. A Powerlink machine must then be demonstrably separated and monitored, like any other industrial automation system.

❓ Frequently asked questions

Is Powerlink open source?

The Powerlink specification is open and may be implemented without licence fees. In addition there is openPOWERLINK, an open-source stack under a BSD licence that can run as a Managing Node and as a Controlled Node. B&R also supplies its own commercial stacks and hardware.

Does the EPSG still exist?

No, the Ethernet POWERLINK Standardization Group was dissolved on 31 March 2023. B&R is its formal successor and has been responsible for Powerlink development, support and certification since April 2023. The specifications are still publicly available.

What is the difference between Powerlink and EtherCAT?

Powerlink has a Managing Node poll each device in turn with separate frames, whereas EtherCAT passes one frame through all slaves and processes it on the way. EtherCAT therefore often achieves shorter cycles but requires dedicated slave chips. Powerlink works with standard Ethernet hardware.

Why does Powerlink use hubs instead of switches?

Within the real-time domain the Managing Node already prevents two devices from transmitting at once, so switching is unnecessary. A repeating hub forwards frames immediately and adds less delay and jitter than a store-and-forward switch. Outside the real-time domain you simply use switches.

How many devices fit in a Powerlink network?

A Powerlink network supports up to 240 nodes: Controlled Nodes get addresses 1 to 239 and the Managing Node has address 240. With multiplexing, less critical nodes can share a time slot so the cycle time stays short even with many devices.

Can Powerlink carry IP traffic?

Yes, Powerlink carries IP traffic in the asynchronous phase of each cycle, in which one node at a time is granted the right to send. Parameterisation, diagnostics and web interfaces can therefore share the same network without disturbing the isochronous process data.

📌 In summary

Powerlink is an open, deterministic Ethernet protocol in which a single Managing Node allocates the right to transmit in every cycle, enabling motion control with cycles below 200 µs on standard Ethernet hardware.

The protocol was introduced by B&R in 2001, is standardised in IEC 61158 and IEC 61784-2, and has been managed by B&R since the EPSG was dissolved in 2023. Its CANopen foundation, openSAFETY and the open-source openPOWERLINK stack make it attractive to machine builders. Because Powerlink has no built-in security, the real-time domain belongs isolated behind a gateway and firewall, just like other industrial Ethernet protocols.