It is becoming increasingly clear that we are entering a new era in systems software. As the age-old tension between structure and performance acquiesces, we can finally venture beyond monolithic systems and explore alternative modularizations better suited to evolution and adaptation. The OASIS project explores the potential of aspects to naturally shape crosscutting system concerns as they grow and change. This paper describes ongoing work to modularize evolving concerns within high-performance state-of-the-art systems software, and outlines some of the major challenges that lie ahead within this domain.