Measurement is the most common method of worst-case execution time estimation in commercial use today. Measurement techniques insert profiling code into the software, recording the end-to-end execution time of sub-systems, functions, or individual blocks of code.

Advantages:

• Measurement provides an initial indication of the execution time performance of software components.
• The measurements are observations of the timing behaviour of the actual hardware.
• Execution time measurements are easy to obtain during testing.

Disadvantages:

• During testing, it is difficult to ensure that the worst-case path has been taken.
• It is also difficult to ensure that the functions and loops on a given path have exhibited their worst-case execution times simultaneously.
• For complex software, running on advanced microprocessors, the longest observed end-to-end execution time is likely to be less than the actual worst-case time.
• Measurements alone are rarely sufficient to provide high levels of confidence in the worst-case timing behaviour of the system.

Alternative methods of WCET estimation include static analysis, and hybrid approaches combining both static analysis and measurement techniques.