Fault characterization is an important part of combinatorial testing which enables the automatic identification of failure-inducing combinations. Up until now, many different algorithms are proposed to compute failure-inducing combinations. However, the only comparisons between different algorithms are done by the algorithms authors themselves who only evaluate few algorithms at a time which complicates comparisons. Therefore, we present a concept and a reference implementation of a comparison infrastructure that allows to evaluate fault characterization algorithms in a comparable manner. In addition, we report on the results of a preliminary comparison using the comparison infrastructure.