Insights into providing dynamic adaptation of operating system policies

Abstract

The DAiSES research project, initiated in September 2004, has allowed us to understand the potential value of and challenges associated with providing dynamic OS policy adaptation. During this period the project has had two major research foci, investigation of process scheduling as a possible target of adaptation and dynamic adaptation of I/O scheduling. The paper revisits the DAiSES methodology, elucidating it in terms of these two adaptation targets. It abstracts the challenges in dynamic adaptation of stateless systems and provides solutions to these challenges, exemplifying them in the context of I/O schedulers.