An abstract model of a distributed data base application has been studied using process based, state based, and queueing theory based methods. The methods supported by graphical notations and/or integrated development environments were found to be easiest to work with. The methods supported by model checkers were the most successful in obtaining relevant information about the application. Applying a number of different methods to study one particular model encourages a problem to be viewed from different angles. This gives complementary information about the model. We report on a variety of problems of the model found through various routes. Our main conclusion is that asking experts to apply different methods and tools at a sufficiently abstract level is effective.