COTS Software Evaluation and Integration issues

This report begins with a survey of COTS evaluation and integration state-of-the-art. Selected parts of the state-of-the-art form the base for our approach to a COTS-based development process. The aim of the process is to identify a combination of COTS products that is optimized for ease of integration and some particular non-functional requirements in the context of current software evaluation and integration state-of-the-art. The process consists of four phases. The first phase project and COTS products characterization characterizes the project requirements and candidate COTS products. Candidate products not satisfying defined selection criteria are screened out. The second phase COTS product measurement measures remaining candidate products according to a defined set of metrics. The evaluation results are compared to a defined set of threshold values, and all candidate products not exceeding these threshold values are screened out. The metrics are given weights indicating that some qualities are of more importance that others. Combinations of products having a high score are ranked over combinations having a low score. The third phase architectural and interface mismatch evaluation compares the architectural assumptions of the remaining candidate products (now called alternative products) to the baseline software architecture to identify architectural and interface mismatches. Based on these mismatches total integration effort is estimated. Product combinations are ranked according to total integration effort, with least integration effort combinations ranked first. The fourth phase optimal integration combination identification combines the quality score ranking and the integration effort ranking to identify a final ranking of product combinations. The top-ranked combination identifies the optimal product combination to be integrated.