Formal verification of safety of interlocking systems and of their configuration on a specific track layout is conceptually an easy task for model checking. Systems that control large railway networks, however, are challenging due to state space explosion problems. A possible way out is to adopt a compositional approach that allows safety of a large system to be deduced from the formal verification of parts in which the system has been properly decomposed. Two different approaches have been proposed in this regard, differing for the decomposition assumptions and for the adopted compositional verification techniques. In this paper we compare the two approaches, discussing the differences, but also showing how the different concepts behind them are essentially equivalent, hence producing comparable benefits.