Software Design Approaches for Mastering Variability in Database Systems.

For decades, database vendors have developed traditional database systems for dierent application domains with highly diering requirements. These systems are extended with additional functionalities to make them applicable for yet another data-driven domain. The database community observed that these \one size ts all" systems provide poor performance for special domains; systems that are tailored for a single domain usually perform better, have smaller memory footprint, and less energy consumption. These advantages do not only originate from dierent requirements, but also from dierences within individual domains, such as using a certain storage device. However, implementing specialized systems means to reimplement large parts of a database system again and again, which is neither feasible for many customers nor ecient in terms of costs and time. To overcome these limitations, we envision applying techniques known from software product lines to database systems in order to provide tailor-made and robust database systems for nearly every application scenario with reasonable eort in cost and time.