A two-step model is proposed and prototyped to assist decision makers selecting n-tiered architectures for Internet applications. The model is developed using goal programming (GP) for the selection process and Monte Carlo simulation for the budgeting process. The GP model explores the impact that software attributes such as scalability, reliability, maintainability, interoperability, understandability, and timeliness have on the selection of Internet application architectures. The simulation uses normal and uniform distributions to model uncertainty in the budgeting process for 14 n-tier architectures. The models are integrated to provide a feedback loop in the selection process. Both conceptual and implementation models are provided.