Software Architecture at a Large Financial Firm

George Fairbanks, Kevin Bierhoff, and Desmond D'Souza

Proceedinngs of OOPSLA 2006, Portland, Oregon, Oct. 22-27, 2006, . To Appear

Online links: PDF

System builders have historically used informal software architecture models to understand options, make choices, and communicate with others. Research into software architecture over the past fifteen years has indicated that more precise architecture models may be beneficial. At a large financial firm, we applied precise software architecture techniques on four software projects and this experience has revealed a number of practical issues. We made the following observations across the projects: 1) Architecture models can be used to bridge gaps between business requirements and technology, 2) A small collection of techniques and a detail knob are practical and useful in a variety of projects, 3) Architecture modeling techniques amplify the skills of the architects, 4) A model of domain concepts and relationships is helpful when building architecture models, and 5) It is difficult to know when to stop adding detail to your architecture model. We believe that these observations motivate future research and can help practitioners make software architecture more effective in practice.

For further information, please visit the home pages of the ABLE research project and Carnegie Mellon University's Composable Systems Group.

Last modified: 8/16/2006. For comments and problems, contact