Multi-level testing constitutes an approach for test reuse across
multiple test levels that questions this strict separation with
respect to specific functions of large embedded systems, namely
those whose functional core is not affected by partitioning or
decomposition. Under these conditions, the integration process
does not impact the functional core, which is implemented entirely
by a single test object at each test level.
The test activities at the different test levels resemble each other
in this context, leading to a degree of commonality across test levels
that reuse approaches can benefit from. In effect, we have observed
that similar or even identical functional test cases are specified,
designed, implemented, executed, and evaluated in parallel at different
test levels for numerous system functions in practice—an
approach that clearly lacks efficiency.