understandable that plant engineers and production managers highly value fault-tolerant network connections. The simplest form of network redundancy is the ring topology. In a ring, network traffic can flow in either direction, providing an alternative path around a break in any single segment of the network.

The proper operation of any Ethernet network demands that there be one and only one path for a network packet to travel. As a simple example of why this is true, consider a network broadcast message traveling around an endless ring. Nothing prevents each broadcast packet to travel endlessly around the ring. Soon the network is flooded with homeless packets and new packets cannot find the opportunity to enter the highway. The solution to this problem is to have an intelligent checkpoint at one point in the ring to prevent this undesirable endless network path. When a break is detected in the Ring path, this intentional block is removed and all network nodes are notified to redirect the traffic as necessary to avoid the road block caused by the break.

SIXNET Real-time Rings™ provide a very fast and highly reliable means to connect Ethernet devices in a Ring. Intelligence in each Ring switch constantly monitors the quality of the link by testing the “Link” status of its connection to its neighbors and by actively sending a test packet around the ring and verifying its successful travel around the ring. These test packets are sent using an urgent priority set in the Ethernet message frame. This ToS (Type of Service) prioritization ensures that these test messages travel unimpeded in a Real-time Ring, even in the face of heavy network traffic, because of their high priority status. This prioritization is a key element in guaranteeing the deterministic performance of a network Ring. Real-time Ring switches have a worst case recovery time of a mere 30 mS plus 5 mS for each switch added to the ring.

Many competitor’s Ring switch products rely only on the Link status to detect a break in the network. While this method finds clean breaks (no connection at all) it fails to detect loss of intelligence within a switch (a failure of the switch itself), degradation in the cabling, or other situation that may be losing message packets. The use of active messaging in SIXNET Real-time Rings™ provides a positive assurance that the Ring is functioning and actually passing valid messages.

SIXNET Real-time Ring™ switches do not rely on a designated master switch to control the Ring. This is a unique feature and a HUGE benefit. A designated master is a unique element on the network. Should it fail, functionality of the Ring cannot be maintained, even by bypassing the failed element. Further, the assignment of a switch to be the master switch requires special configuration before that switch can be placed in service. This increases the required skill level of the installer and will most likely increase the time required to replace a failed switch. By its very nature, unique and critical elements such as a designated master is contrary to the very nature of best redundancy logistics.

Most Ring Switch product offerings are based upon proprietary modifications of the Spanning Tree algorithm. Spanning Trees offer great flexibility in network topology but when they are connected in a simple ring, still carry the burden of their complexity and the resulting inefficiencies. Spanning trees were designed for non-real-time office networks. They simply do not guarantee network performance. Even Rapid Spanning Tree implementations can occasionally take more than a second to recover from a network break. Beware of Ring solutions that require Managed Switches (or are delivered as a feature of Managed Switches). These schemes utilize modified Spanning Tree software.

Please also note that Managed Switches, while more capable overall than Real-time Ring Switches, carry the burdens of their complexity, including higher cost and higher skill level required to configure and maintain them. Ask yourself: “Do I need managed switches in this application?”

SIXNET holds a patent on the capability to report network status using Modbus or other common industrial protocols. This ingenious innovation enables you to monitor network status and find breaks in the network path using simple Modbus registers. Network “alarms” can be detected in PLCs and HMI devices in the same manner as any other alarm condition. This capability is an interesting alternative to SNMP (found only in Managed Switches), particularly in automation systems that must be maintained by plant electricians without the assistance of network specialists.

As an aid to system designers, we have provided some questions to consider before selecting a Ring Switch solution for your real-time systems. If you require further advice, please feel free to contact the experts at SIXNET. support@sixnetio.com

Important Questions to Ask