Risk/return management has not only evolved as one of the key success factors for enterprises especially in the financial services industry, but is in the times of the economic crisis initiated by the financial markets crucial for the survival of a company. It demands powerful and at the same time flexible computational resources making it an almost ideal application for service-oriented computing concepts. An essential characteristic of service-oriented infrastructures is that computational resources can be accessed on demand and paid per use. Taking the estimation of covariances for a portfolio of risky investment objects as an example, we propose quantification for the economic value of fast risk/return management calculations. Our model then compares the cost structures of serviceoriented infrastructures and dedicated systems in this domain. On this basis, we can determine under which circumstances the one or the other architectural strategy is superior.