Plug-in hybrid electric vehicles (PHEV) are becoming gradually more attractive than internal combustion engine vehicles, even though the current electrical grid is not potentially able to support the required power demand increase to introduce charging stations. Acknowledging that design and development of charging stations has crucial importance, this paper introduces a candidate PHEV charging station architecture, along with a quantitative stochastic model, that allows us to analyze the performance of the system by using arguments from queuing theory and economics. A relevant component of the proposed architecture is the capability of the charging stations to store excess power obtained from the grid. The goal is to design a general architecture which will be able to sustain grid stability, while providing a required level of quality of service; and to describe a general methodology to analyze the performance of such stations with respect to the traffic characteristics, energy storage size, pricing and cost parameters. Our results indicate that significant gains in net cost/profit and useful insights can be made with the right choice of storage size. Such considerations are crucial in this early stage of designing the smart grid and charging stations of the future.