Some embedded systems are . By that, we mean that they are in places where they do a very important job and if they fail, it may have implications. Embedded systems in planes, for example, are often critical. These types of embedded systems need a different design approach to the embedded systems found in washing machines, for example.

1) They need to run . When we looked at the alarm clock, one way to design the program was using interrupts. If the backlight button was pressed, then an interrupt occurred that resulted in a procedure being called. If that procedure was called within 3 milliseconds one time and within 20 milliseconds the next time it wouldn't really matter. With critical systems, it a lot. It is very important to be able to predict with the reaction times to interrupts and the time that some code will run.

2) They need to run . It is important that a mission-critical system will work whatever happens! To achieve this, critical embedded systems have many systems built in. They may have a second or even processor, for example, ready to step in and take over if the main processor fails. They may have methods that allow data to be read in twice using different sensors and checked against each other. They may have sub-systems ready to take action monitoring the main systems. If necessary, a sub-system can, for example, alert a pilot to a problem. They may have software that can attempt to from faults. Programming languages such as are designed for this.

Critical systems are . They have a lot of hardware and software that may never be needed. Although any computerised system can never be 100% reliable, critical embedded systems attempt to reduce the risk to acceptable levels.