The Research On Task-independent Scheduling Model And Its Implementation In Extensible OS

All the scheduling models and frameworks of traditional operating systems are designed based on the special requirement of different tasks. Although supplying the specified tasks with high efficiency, they sacrifice the performance of the others. The main reason is that the task-specified scheduling logic is a component of the kernel and is tightly connected with other codes. Thus, the whole system is lack of generality, portability and extensibility.As the shortcomings of the scheduling model, this paper presents a kind of brand-new scheduling framework, according to the requirement of the extensible operating system in orthogonal architecture. The framework is a task-independent and fair scheduling model, which is divided into two layers: task-dependent layer and task-independent layer. The task-dependent layer brings forward the corresponding requests. After the admission control, task-independent layer guarantees the specified scheduling requirement by scheduling processes. While keeping a base kernel, the code of task-dependent-scheduling is downloaded into the kernel to provide the performance of scheduling.At first, the paper introduces the total framework of task-independent scheduling model. And from two layers of task-independent and task-dependent, we discuss the design and implementation of the scheduling model. Then, we analyze the functions of admission control and its process of working in the model. Moreover, we explore the scheduling model deeply on the aspects of extensibility, fairness, safety, guarantee of QoS and efficiency, etc. At last, we carry out the further work in progress.