Continuously increasing global container trade and pressure from a limited number of large shipping companies are enforcing the need for efficient container terminals. By using internal material handling resources efficiently, transfer times and operating costs are reduced. We focus our study on container terminals using straddle carriers for transportation and storage operations. We assume that straddle carriers are shared among maritime and inland transport modes (truck, train, barge). The problem is thus to decide how many resources to allocate to each transport mode in order to minimize vehicle (vessel, truck, train, barge) delays. We present a mixed integer linear programming model, based on a network flow representation, to solve this allocation problem. The modular structure of the model enables us to represent different container terminals, transport modes and service strategies. We present parts of our model and exemplary applications for a terminal at the Grand Port Maritime de Marseille in France.