top logo netmoduleheader backgroundclaim netmdule: we speak embedded!

Applying PRP and HSR Protocol for Redundant Industrial Ethernet

IEC62439/IEC62439 provides redundancy without packet loss. This is a crucial advantage for time-critical applications. - Applying PRP and HSR Protocol for Redundant Industrial Ethernet

The IEC62439 standard describes methods to implement Ethernet based network redundancy without packet loss. This is a crucial advantage for time- and safety-critical applications.

Ethernet is the predominant communication protocol in the office world, datacenters and for higher level communication in the automation world. Deploying the same protocol for communication between controllers and field devices leads to a vertical integration of field communication together with Manufacturing Execution Systems (MES).
However, deploying the Ethernet protocol for fault tolerant networks with a high availability and short reaction time leads to additional requirements. Real time systems such as robot control, synchronized drives and substation automation require reaction times in the millisecond area.
On the other hand, Ehernet technology’s broadcast nature does not permit physical loops and therefore effectively forbids redundant communications paths.
The IEC62439-3 standard offers solutions to address this challenge. The Parallel Redundancy Protocol (PRP) and High availability Seamless Redundancy (HSR) protocol are suited for applications that demand high availability and low latency and thus gain momentum in the SCADA area. The PRP and HSR protocols provide a seamless fail-over. This means that no single packet is lost in case of a single network failure. PRP and HSR thus provide a clear advantage over other redundancy mechanisms that require a failover time of several 100ms such as RSTP, MRP (IEC62439-2) or HIPER-Ring.
Redundancy is provided on layer 2. Thus, higher layer industrial Ethernet protocols run seamlessly with PRP and HSR.
A good example for the shift towards Ethernet in mission critical applications is substation automation: The IEC61850 standard for substation communication is based on Ethernet as core technology. The standard proposes the use of HSR and PRP for media redundancy. Due to its high flexibility, cost effectiveness and high performance, the standard has gained significant traction and utilities start to quickly replace legacy and expensive proprietary technologies with Ethernet technology.

Properties of high-availability technologies

Software drivers for the PRP can be installed on any platform supporting standard operating systems such as Windows and Linux or on dedicated RTOS such as VxWorks. The throughput is sufficient for most fast Ethernet applications. 
more >

For more time-critical applications, the link redundancy layer can be implemented in FPGA technology. This allows for wire speed operation and fast cut-through HSR operation. In addition, the Precision Time Protocol (PTP) as defined in IEEE 1588 standard can provide robust synchronization in the sub-microsecond range. 

Parallel Redundancy Protocol (PRP), IEC62439-3 clause 4
Double Attached Nodes (DANs) with critical functionality have two network ports connected to two physical independent networks (LAN A and LAN B). Standard networking equipment ((switches, Ethernet NICs)) can be used to build the two networks.
DANs transmit packets on both interfaces. Under normal conditions, a destination DAN receives all packets in parallel on both interfaces and discards duplicates.
The protocol supports also Single Attached Nodes (SANs) in a PRP network. They only connect to one LAN and do not make use of the network redundancy.
A Redundancy Box (RedBox) can be used to connect one or more conventional nodes (with only one network interface) to both LANs of a PRP network. Nodes behind the RedBox virtually attach to both networks and are thus called Virtual Dual Attached Node (VDAN).

IEC62439 PRP

PRP provides redundancy on OSI layer 2 and both network ports of a DAN share the same MAC and IP address. The mechanism is completely transparent for higher layer network protocols and applications.
more (German only)>

High availability Seamless Redundancy (HSR), IEC62439-3 clause 5
Each node on the network has two network ports and is directly inserted in a network with ring topology. Packets are transmitted on both network interfaces and thus transmitted in both directions of the ring at the same time. Intermediate nodes forward the frames in the ring. The receiver removes unicast-frames from the ring, the sender removes multi-cast und broad-cast frames. In error-free conditions, the destination node receives two copies of the frame and discards the duplicate.
Since forwarding delays of all intermediate nodes sum up, device internal delays should be kept as small as possible. This calls for a cut-through forwarding mechanism implemented in hardware. HSR supports this by using a header structure that is optimized for hardware parsing.

Two HSR rings can be coupled using a QuadBox. For redundant coupling, two QuadBoxes are required. PRP networks and HSR rings are coupled using a RedBox.

IEC62439 HSR

The ring topology of HSR provides several advantages:

  • All nodes participate in packet forwarding, no additional networking equipment is required
  • Deployable where cabling with ring topology is inherent (e.g. infrastructure)

HSR implements the redundancy on OSI layer 2 and is transparent for upper layer protocols and applications. It can be used as a redundancy mechanism for other industrial Ethernet protocols lacking redundancy such as EtherNet/IP, ModBus/TCP and others.

Stefan Leuenberger  Netmodule

Stefan Leuenberger

+41 31 985 25 10

Send E-Mail to Stefan Leuenberger

Patrick Zysset  Netmodule

Patrick Zysset

+41 31 985 25 10

Send E-Mail to Patrick Zysset