The many faces of multi-level real-time scheduling

Many real-time scheduling algorithms use simplistic sets of assumptions and this limits their applicability in practice. Often, a large real-time system has multiple scheduling algorithms and multi-level scheduling algorithms. Multiple algorithms arise when it is possible to partition the system into subsystems, each with their own algorithm. While this is a valuable approach, in this paper we focus on the issue of multi-level scheduling. Multi-level scheduling arises for various important reasons. In this invited presentation we discuss multi-level scheduling examples from a process/thread model, local/distributed scheduling, manufacturing, and multimedia.