In the context of the benchmark we are proposing, we
must characterize the relevance of the security mechanisms
from a usefulness perspective. In fact, we want the
benchmark to be as accurate as possible, but also as
representative as possible, which means that taking too much
environment assumptions into account would probably
makes it less useful for a large number of scenarios (i.e.,
scenarios where those assumptions are not valid). In this
issue, our approach is to try to determine classes of
mechanisms that can justifiably be considered with the same
relative importance. To this end, we use a relative ordering
established through the consensual average opinion of
several security experts. The relative order is used to
determine four groups of mechanisms, which we consider
equivalent. We believe that this approach provides a reliable,
yet flexible, way of accounting for the importance of the
security mechanisms.