In order to push Electric Vehicles (EVs) into the mainstream, the wide deployment of charging stations that can serve multiple classes of customers (e.g. fast charge, slow charge etc.) and provide a certain level of Quality of Service (QoS) is required. However, the operation of the power grid becoming more strenuous due to the addition of new large loads represented by EVs. Hence in this paper we propose a control and resource provisioning framework that can alleviate the strain on the power grid. We propose two design problems; first one considers a charging station located in a big metropolitan with a large and highly stochastic EV demand. For this case, we propose a pricing based control mechanism to maximize the total aggregated utility by controlling the arrival rates. Second case provides a capacity planning framework for stations located in small cities where arrival rates can be obtained via profiling studies. At each model, station draws a constant power from the grid and provides QoS guarantees, namely blocking probability, to each class. Hence total stochastic demand is replaced with a deterministic one, by sacrificing to reject a very few percentage of customers. Our results indicate that significant gains can be obtained with the proposed model.