We model and analyze the process of passengers boarding an airplane. We show how the model yields closed-form estimates for the expected boarding time in many cases of interest. Comparison of our computations with previous work, based on discrete event simulations, shows a high degree of agreement. Analysis of the model reveals a clear link between the efficiency of various airline boarding policies and a congestion parameter which is related to interior airplane design parameters, such as distance between rows. In particular, as congestion increases, random boarding becomes more attractive among row based policies.