Metasys BACnet/IP equipment controllers may be physically connected to your building automation network in multiple ways. In this document, we refer to the topology options as star, chain, ring, and mixed. Most networks use combinations of these topologies. Chain and ring topologies require that each device have two Ethernet ports. The IP controller has an embedded Ethernet switch which regenerates the signal when passing Ethernet packets between its ports. Therefore, the distance between IP controllers is limited by the 100-meter limit imposed by Ethernet. Some controllers can be more than 100 meters from the switch if the path back to the switch is hops of 100 meters or less.
In a star network, each Metasys BACnet/IP equipment controller is connected directly to a main switch. This reduces the possibility of network failure but requires more wiring and labor to install, both for the initial controllers and for additional controllers added later. Because of the 100 meter limit imposed by Ethernet, in star topology all controllers must be within 100 meters of the switch. Using a star topology also increases the number of required switch ports, and more ports means more network switching hardware. If the controllers are connected to the customer's IT network, on-going IT administration costs may be incurred for these ports.
In chain and ring topologies, each controller is required to have two Ethernet ports. To chain Metasys BACnet/IP equipment controllers, connect an Ethernet cable from the switch to the ETH1 port on the controller. Add an additional controller by connecting an Ethernet cable from the ETH2 port on the connected controller to the ETH1 port on the new controller. For a chain, one and only one Ethernet cable terminates at the network switch. The ETH1 and ETH2 ports are interchangeable, and cables can be connected to either port. Also, new controllers can be added to the middle of the chain or ring. While there is no maximum length of a chain, there is a reduction in reliability as more devices are added to the chain. Up to 100 devices can be added to a ring, but reductions in reliability should still be considered when determining the size of a ring. For more information on comparing the reliability of chains and rings, as well as factors to consider when adding a device to a chain or ring see section Network reliability and controller availability.
A ring network is like a chain, except one additional network cable is used to connect the free end of the chain back to the Ethernet switch. Accidental rings are a problem for Ethernet networks because it is possible for network traffic to circulate endlessly around a ring. On the other hand, a properly managed ring can provide redundancy. The IP controllers from Johnson Controls have two Ethernet ports so that they can be connected in rings. They also have software support for Media Redundancy Protocol (MRP). MRP is a data network protocol standardized by the International Electrotechnical Commission as IEC 62439-2. MRP allows a large chain of Ethernet devices to overcome any single communication failure, with a recovery time faster than can be detected in a BACnet system. This topology is called a ring because the chain of controllers is virtually closed by a software component in the switch called a Media Redundancy Manager. The software works in cooperation with a Media Redundancy Client in the controllers. Together, the software components assure that the physical ring always behaves like one or two logical chains. To use MRP, configure two ports of a switch as ring ports. Note that not all switches support MRP.
The chain and ring topologies offer a more economical method of connecting controllers to the network than a star topology. First, because controllers can connect to each other, connecting all of them requires fewer switch ports. Second, because controllers are closer to each other than to the network switch, connecting all of them requires less Ethernet cable. Third, they offer a simple means of connecting controllers that are more than 100 meters from the switch. For Fast Ethernet, the most common kind of Ethernet and the kind used for Metasys controllers, the length of a single cable is limited to 100 meters. However, the IP controllers have an internal Ethernet switch which regenerates the signal when passing Ethernet packets between its ports, meaning a single chain or ring can be much longer than 100 meters.
The difference between chain and ring topologies are related to reliability. For chain, a controller failure or Ethernet link failure results in a loss in network connectivity for all controllers in the chain which are on the far side of the failure from the switch. Ring topology is better because it provides protection from a single point of failure, either failure of the link or of the controller. Only multiple failures result in a loss in network connectivity for other controllers. For star, the controller loses network connectivity only when there is a failure of the Ethernet link between the controller and the switch.
