Table of Contents
Advertisement
P24xM
5
REDUNDANT ETHERNET COMMUNICATION
Redundancy is required where a single point of failure cannot be tolerated. It is required in critical applications
such as substation automation. Redundancy acts as an insurance policy, providing an alternative route if one
route fails.
Ethernet communication redundancy is available for most General Electric products, using a Redundant Ethernet
facility. This is an in-built Network Interface Card (NIC), which incorporates an integrated Ethernet switch. The
interface provides two Ethernet transmitter/receiver pairs, both of which are for the same physical medium (two
copper, or two fibre). In addition to the two Ethernet transmitter/receiver pairs, the REB provides link activity
indication.
Both industry standard PRP (Parallel Redundancy Protocol) and HSR (High-availability Seamless Redundancy)
protocols are available by ordering option.
5.1
SUPPORTED PROTOCOLS
One of the key requirements of substation redundant communications is "bumpless" redundancy. This means the
ability to transfer from one communication path to another without noticeable consequences. Standard protocols
of the time could not meet the demanding requirements of network availability for substation automation
solutions. Switch-over times were unacceptably long. For this reason, companies developed proprietary protocols.
More recently, however, standard protocols, which support bumpless redundancy (namely PRP and HSR) have
been developed and ratified.
P40Agile version 56 onwards supports redundant Ethernet. Variants for each of the following protocols are
available:
PRP (Parallel Redundancy Protocol)
●
HSR (High-availability Seamless Redundancy)
●
PRP and HSR are open standards, so their implementation is compatible with any standard PRP or HSR device
respectively. PRP provides "bumpless" redundancy.
Note:
The protocol you require must be selected at the time of ordering.
5.2
PARALLEL REDUNDANCY PROTOCOL
PRP (Parallel Reundancy Protocol) is defined in IEC 62439-3. PRP provides bumpless redundancy and meets the
most demanding needs of substation automation. The PRP implementation of the REB is compatible with any
standard PRP device.
PRP uses two independent Ethernet networks operating in parallel. PRP systems are designed so that there should
be no common point of failure between the two networks, so the networks have independent power sources and
are not connected together directly.
Devices designed for PRP applications have two ports attached to two separate networks and are called Doubly
Attached Nodes (DAN). A DAN has two ports, one MAC address and one IP address.
The sending node replicates each frame and transmits them over both networks. The receiving node processes the
frame that arrives first and discards the duplicate. Therefore there is no distinction between the working and
backup path. The receiving node checks that all frames arrive in sequence and that frames are correctly received
on both ports.
Devices such as printers that have a single Ethernet port can be connected to either of the networks but will not
directly benefit from the PRP principles. Such devices are called Singly Attached Nodes (SAN). For devices with a
single Ethernet port that need to connect to both LANs, this can be achieved by employing Ethernet Redundancy
P24xM-TM-EN-2.1
Chapter 16 - Communications
311
Advertisement
Table of Contents
Troubleshooting
