Metadata-driven design of integrated environments for software performance validation

Lifecycle validation of the performance of software products (i.e., the prediction of the product ability to satisfy the user performance requirements) encompasses the production of performance models from CASE documents. The model production activity is a critical, time-consuming and error-prone activity so that lifecycle validation is still not widely accepted and applied. The reason is twofold: the lack of methods for the automatic derivation of software performance models from CASE documents and the lack of environments that implement and integrate such methods. A number of methods for the automatic derivation of software performance models from CASE documents has been already proposed in literature, without however solving the automation problem. This paper instead faces up to such problem, by introducing an integrated and standards-based environment for the automatic derivation and evaluation of queueing-based performance models. The environment is based on the use of standards for metadata exchange (MOF, XMI), to ease the integration of the most common UML-based CASE tools, thus enabling software designers to smoothly introduce performance validation activities into their best development practices.