Is it worth spending 1 million EUR on a new software architecture? The new software architecture might improve the software performance – but to which extend? We help in identifying cost-efficient software design alternatives. Using the innovative Palladio simulation approach, we can recommend valuable design decisions based on science instead of gut feeling.
The storage subsystem of the IBM z-Series is a very complex, virtualised, multi-layer software system. In a project with IBM in Böblingen, Germany, we successfully modelled and predicted the performance of different architecture design alternatives of the storage subsystem. This enabled IBM to compare the architectural design alternatives without requiring to implement at least prototypes of all promising alternatives. The model-level prediction capabilities of Palladio helped to avoid wasting implementation efforts for poorly performing design alternatives. The project could successfully show that Palladio helps selecting the right design alternatives while at the same time lowering the effort for the evaluation of quality attributes of software architectures. Furthermore, Palladio helped identifying potential bottlenecks and enabled sizing analysis for different load scenarios.
Overview on further reference projects and industrial partners.