Industrial Ethernet Redundancy Basic

Ethernet redundancy has be adapted in the industrial automation or industrial networking in the recent years. It enhances the reliability of the data communication for the complex and large networks in the industrial world. There are several different topologies to form the redundancy network. Upon the importance and urgency, engineers can choose one of topologies below for their Ethernet network.

Technologies

  1. Mesh network
    The easiest way to create the redundancy network is connecting every single node to each other. This is called Mesh network. In this topology, some nodes are connected through a single hop (node), and some are connected with more than one hop. Data traveling on the mesh network is automatically configured to reach the destination by taking the shortest route. If any one of the hop (node) failed, each node will find an alternate route automatically to transfer the data. It is highly impossible that all the nodes within a mesh network fail at the same time. The mesh network is highly reliable for data transmission; however, it has a much higher communication overhead (due to the possible multi hops to route the data transmission) and the much higher cabling cost which is required to connect all the nodes to each other.

  2. STP (Spanning Tree Protocol)
    STP is one of the popular Ethernet redundancy protocols which is defined by the IEEE. It reduces the cabling cost compared to the Mesh network since the nodes no longer need to be connected to each other. The trade off is the recovery time is slow and unsuitable for the industrial applications.

    The idea of STP is to eliminate loops (such as in the mesh topology) in a network by cutting the network into a loop-free tree shape. It helps a network achieve link redundancy and path optimization. STP uses an algorithm to find redundant links in a network and allows certain paths as backup paths to prevent looping. When the primary connection fails (ex: switch 5 to switch 3 on the diagram), STP then automatically re-arranges the connections by activating the backup paths to forward data.

    With STP, the recovery is slow and usually takes 15 seconds after the spanning tree is established. The whole recovery process can take up to 50 seconds.
     
  3. RSTP (Rapid Spanning Tree Protocol)
    Similar as STP, the RSTP is designed with the same topology with the improvement of STP slow recovery time. RSTP shortens the recovery time up to 15 seconds, but is still not good enough for the mission critical applications.

    STP & RSTP are open standards that many Ethernet switch manufacturers have implemented these features in their managed switch products. The RSTP is very helpful with faster self-healing time in an enterprise network which most of the tasks don’t require real-time connection. A few seconds delay usually doesn’t cost any damage and is acceptable. For other mission critical application such as military and industrial networks, a much faster recovery is required for safe operation and to prevent expensive system downtime.
     
  4. Ring
    The Ring redundancy is the most common among today’s industrial Ethernet or mission critical networks. This solution is more cost-effective than a mesh network, and has a much faster recovery time than STP/RSTP. The Ring redundancy technologies guarantee recovery time of a few milliseconds (less than 20 milliseconds) by the switch providers. The Ring ensures a non-stop operation of network and is ideal for the mission critical applications. For example, within a ring network, if any segment of the network is disconnected, the network system will recover in less than 20 ms by activating the backup path in a ring.

    The drawback of the ring protocol is that it is a proprietary protocol, and it is different from manufacture to manufacture. The interoperability is a challenge due to this reason. In addition, ring redundancy only supports limited rings in a single switch.

All the above topologies such as mesh, STP/RSTP and ring are designed for the infrastructure level of a network. In another word, it is designed for the backbone communication. What can we do if we want to extend the network redundancy all the way to the end devices? The answer is Chain.

Chain

Same as the ring protocol, chain is also a proprietary protocol. Chain works by connecting several Ethernet switches together to form a daisy-chain, where a “Head” switch and a “Tail” switch (the edge switches at the both end of the chain) are configured. Simply link the 2 ends of the chain to an Ethernet network and the setup is done. It is shown as the illustration below.

Once a chain fails, the blocked path (link between the “Tail” and the outside network) will rapidly be activated within 20 milliseconds. It ensure the Ethernet network remains available almost constantly.

Summary

The mission critical applications such as military and industrial network require reliable, redundant network system to ensure non-stop network operations. Even if there are many different existing redundant protocols that can achieve the similar purpose, only the ring and chain topologies can self-heal within few milliseconds and fully satisfy the requirements.

Click here for products that support Ethernet redundancy.

Related Articles