The benefits of modular design
Suppose we are required to produce a range of different clocks, with various forms of display (Figure 25.2).
Some of the clocks may have different features (for example, the ability to set an alarm), but the key tasks are the same in all cases: to keep accurate track of the time and display this information on a display.
In some circumstances, it may be useful to distribute the application over two modules, each with a separate microcontroller. The first module would deal with the basic timekeeping and time adjustment facilities; the second module would provide support for the different displays, such as an LCD driver or a stepper motor. This approach may provide economic benefits, since it allows us to produce many thou- sands of the basic timekeeping modules at low cost. We can then produce different displays, as required, to match the needs of particular customers.
This type of modular approach is very common in the automotive industry where increasing numbers of microcontroller-based modules are used in new vehicle designs.
Consider another example. Suppose we have a data-acquisition system with a single processor and a number of distributed (simple) sensors (Figure 25.3).
In this arrangement, if the cable to (say) Sensor 1 is damaged, then no data will be obtained from this sensor until the link is repaired; worse, if an inappropriate data representation has been used, the acquisition system may not even be aware that the link has been damaged.
Consider now an alternative solution using ‘intelligent’ sensors (Figure 25.4).
In this version of the system, Sensor 1 is (we assume) in very close proximity to a microcontroller (‘MCU A’); together, these two components make up our ‘intelligent’ sensor. Communication between the intelligent sensor and the main acquisition
This type of ‘intelligent’ node behaviour can be very useful in many circumstances. For example, in the A310 Airbus, the slat and flap control computers form an ‘intelligent’ actuator subsystem. If an error is detected during landing, the wings are set to a safe state and then the actuator subsystem shuts itself down (Burns and Wellings, 1997, p. 102).
As we will see in the remaining chapters in Part F, most S-C schedulers support the creation of backup nodes, which may be made ‘intelligent’ if this is required.