About Me     Research     CV     Publications         Professional     Personal

Real-Time Systems

A real-time system is is defined as a system that has both :
  • Logical, and
  • Temporal correctness.
The temporal correctness is defined in the form of a constraint, usually termed as deadline. The usefulness of results produced by the system drops significantly, at times, to zero, on the passing of this constraint.

Real-time systems are broadly classified into the following categories :
  • Hard : The usefulness of results on the passing of the contraint falls sharply and missing such constraints could lead to catastrophic effects to the system, then user, the environment, or all of them. Eg. : ABS system in cars, safety control of nuclear reactors, etc.
  • Soft : The usefulness of results falls gradually, on the passing of the constraint, and a few missed deadlines can be tolerated by the system, without catastrophic results. Eg. : MPEG decoding of streaming video.
A real-time system is defined as a collection of tasks, each of which is defined as : {phi, p, e, d}, where :
  • phi : phase
  • p : period
  • e : worst-case execution time of the task - the guaranteed upper bound on execution time of the task, for all possible inputs.
  • d : temporal constraint, deadline.
Real-time systems theory reasons about the schedulability of task sets i.e., offline schedulability tests, which can determine if all deadlines of a set of tasks can be met. Task parameters have to be known beforehand, i.e., parameters such as period of each task, the worst-case execution time (WCET), and so on. Periods of tasks are determined from the operating environment, such as temporal constraints on sensors, actuators and other parts of the system. Determining the WCET for tasks is a non-trivial effort due to software complexity, non-determinism of inputs and hardware complexity with unpredictable execution behavior.
<< Home