Depending on the combination of various WiCAN hardware, the network assumes an appropriate mode of operation. That mode provides a compromise between transparent communication and the latency + bandwidth constraits of the wireless medium.
Point to Point bridge
Two bridges are in association with each other. Assume bridges 'A' and 'B' in the illustration above. All other devices in the image are non-existent. Here CAN messages received by 'A' are wirelessly tranmitted to 'B', which sends the messages out of it's CAN port. And vice versa.
(Future firmware options might provide message filtering capabilities to forward only traffic that is intended for devices that are known to be behind the bridge.)
Multiple bridges are associated with each other. So illustration above, but without 'Ambient wireless' and without 'wireless DMM'. Now, CAN messages received by one bridge are wirelessly tranmitted to all others, each of which sends the messages out of it's CAN port. The concentrator bridge coordinates the network and maintains a multicast list that contains up to 12 devices (including itself). This list is transmitted to all the devices whenever a fresh device is associated with the network. This mode of operation reduces available bandwidth proportionally to the N-1 number of bridges in the network because the multicast transmissions on 802.15.4 are really nothing more than multiple unicast messages, one to each receipient.
P2P Sensor network
One bridge is programmed to be the coordinator for a network of WiCAN sensor devices. From above illustration, assume 'brige B' and bridge 'C' are non-existing. This coordinator maintains a list of all associated sensor devices and their CANopen® node ID. Because those sensors are typically sleeping during the SYNC interval, messages for these devices are stored on the concentrator until after the next SYNC message. After that SYNC, all pending messages are sent to the sensor devices. CAN messages that do not match with the bridge or one of it's associated devices are discarded.
Sensors + star bridge
One bridge is programmed to be the coordinator for a network that consists of WiCAN sensor devices and one or more other bridge devices. Now the multicast list of bridge devices is also propagated to the sensor devices. Those sensor devices send their data using multicast to all the bridge devices in the network.
When using a lot of sensor devices, it's probably better to create two networks; one P2P sensor network to handle the sensor devices and a separate network with bridges to connect two or more CAN segments.
In order to mitigate the differences between the real-time broadcast nature of the physical CAN-bus with it's guaranteed delivery, and the much less ideal nature of radio communication, we have developped some mechanisms that help to overcome those differences:
All devices have two transmit FIFO's to queue outbound CAN messages. One FIFO is for high priority messages with a CAN ID ≤ 0x100. Those messages are send regardless of content of the regular FIFO and regardless of the SYNC interval pause. The normal FIFO is only processed when the priority FIFO is empty and no pause is required in anticipation of an incoming SYNC message.
WiCAN sensor devices are typically sleeping between SYNC pulses to reduce their power consumption. During this sleep, the devices cannot send or receive wireless data. To overcome this, the concentrator bridge maintains a list of associated devices and stores CAN messages intended for those devices in a separate FIFO. After the SYNC message has been send, and all devices are known to be awake, this contents of this FIFO is processed and send to the devices.
CANopen SYNC PLL
Many sensor devices rely on the CANopen® SYNC message to synchronise their sampling rates. In order to get this SYNC message simutaneous to all devices, this message is sent as an unconfirmed broadcast message. In increase the chance of succesfull reception by all devices, all devices maintain a interval timer that is locked to the SYNC interval. All these devices maintain a short period of radio silence around the expected arrival of the next SYNC message.
Additionally, this SYNC interval is used to corodinate the sleep interval of the battery powered sensor devices.
CANopen NMT & LSS regeneration
In CANopen®, network management is performed via common CAN identifiers which are received by all devices addressing of individual devices is done using parameters in te data field of the message. When a bridge receives such messages and has sleeping end devices associated, these messages are not only processed as usual, but also saved in a separate indirect FIFO in order to be broadcast to all WiCAN sensor devices after they have woken up.