To navigation
To content area



Redundancy protocols are used in order to improve the availability of a network. Such protocols typically perform a reconfiguration of the network after a failure of a node or a link, i.e. a new network topology is computed and activated. Applications cannot communicate during this reconfiguration for some 100 ms up to a few seconds, depending on the type of protocol and on size of the network. Demanding applications in the field of automation, transportation, and power distribution require mechanisms free of interruption, i.e. qualified as "with zero switchover time" or "bumpless").

An approach to realize bumpless redundancy is to transmit multiple copies of a frame over different independent paths at the same time. The receiver processes the copy arriving first and discards the duplicates. The "High-availability Seamless Redundancy" (HSR) Protocol implements – similarly as the Parallel Redundancy Protocol (PRP) – this strategy, but in a ring. A ring provides for each communication association two disjoint paths.


The concept of HSR - For a larger view in new window


The concept of HSR comprises the following network elements and options (see figure)

  • A node integrated directly in the ring is denoted as a Double Attached Node implementing HSR (DANH). It is a device with an integrated three port HSR switch (two external and an internal port). A frame to be sent is duplicated and then transmitted in both directions of the ring. Intermediate nodes forward the frames. Unicast frames are removed from the ring by the receiver,   multicast and broadcast frames by the sender. The receiver processes the first frame received and discards duplicates.
  • HSR frames are labelled by a header extension which allows to identify them and facilitates the detection of duplicates.
  • Non-HSR-capable devices can be integrated into the network by means of a Redundancy Box (RedBox). The RedBox provides the HSR functions on behalf of the nodes connected to it. Thanks to this proxy functions nodes beyond a RedBox appear like double attached nodes an are therefore called Virtual Doubly Attached Nodes (VDAN).
  • HSR rings may be interconnected by Quadboxes. A redundant coupling requires two Quadboxes.
  • Networks running the Parallel Redundancy Protocol (PRP) can be connected to HSR networks by means of RedBoxes. A PRP end node is denoted as Double Attached Node implementing PRP (DANP).

HSR RedBox - For a larger view in new window

ZHAW’s prototype of a RedBox: LAN A and B are the connectors to the ring while LAN I (Interlink) connects a non-HSR-capable part of the network.


ZHAW School of Engineering
Institute of Embedded Systems
Technikumstr. 9
CH-8401 Winterthur

Tel. +41 58 934 75 25
Fax: +41 58 935 75 25
E-Mail: info.ines(at)