Support for real-time computing within general purpose operating systems-supporting co-resident operating systems

Abstract

Distributed multimedia applications are typical of a new class of workstation applications that require realtime communication and computation services to be effective. Unfortunately, there remains a wide gap between the development of real-time computing technology in the research community and the deployment of real-time solutions in commercial systems. In this work we explore technology for allowing two operating systems, a general purpose operating system and a predictable real-time kernel, to co-exist on the same hardware. We discuss the problems of multiplexing shared devices and partitioning shared data structures to accommodate two operating systems, and present a CPU executive that allows the IBM Microkernel (a derivative of the Mach microkernel) with an OSF/1 server to co-exist with a simple real-time kernel we have built. We also extend the traditional theory of scheduling periodic tasks on a uniprocessor to accommodate the case where a real-time kernel is allocated only a fraction of the total CPU capacity. . IntroductionRecent advances in high-speed networks, workstation processors, and digital audio and video acquisition/compression/display hardware have led to tremendous interest in the problems of supporting the real-time computation and communication requirements of distributed multimedia applications [1][2][3][4][5][6][7]. However, although much research has been performed on the design of real-time operating systems, and much is on-going in the distributed multimedia domain, the vast majority of computers in use today use operating systems that provide no support for real-time computing beyond allowing tasks to manipulate their execution priority. Operating systems that (1) allow programs to specify their real-time performance requirements and (2) guarantee these requirements are met, have been, and are still, largely confined to academic and industrial research laboratories. To date there have been two identifiable approaches to marrying real-time and non-real-time technologies. First, there have been several attempts to add real-time features such as periodic tasks and priority-inversion-free interprocess communication mechanisms into existing general purpose operating systems such as Unix and Mach [1,3,4]. The second approach is to design a general purpose but dynamically configurable operating system kernel that is capable of accommodating user or application specific process, scheduling, and memory management modules [11]. Here we investigate a third approach: supporting coresident operating systems. Our interest lies in developing a methodology for allowing existing, well-understood realtime systems technology to co-exist with commercial, general purpose operating systems to support applications such as desktop multimedia conferencing. Thus our emphasis is not so much on the development of new realtime systems technology per se, but rather on developing a vehicle for the incorporation and use of existing technology in conventional desktop operating systems.Our approach is to pa...

Using Hierarchical Scheduling to Support Soft Real-Time Applications in General-Purpose Operating Systems

“…The kernel scheduler itself in a general-purpose operating system can be viewed as a hierarchical scheduler: a fixed-priority scheduler implemented in hardware schedules interrupts at the highest overall priority, a (usually) FIFO scheduler runs low-level kernel tasks at the middle priority, and application threads are run at the lowest priority using what is usually thought of as "the scheduler." Furthermore, when the co-resident operating system approach to real-time is used [11,93], the kernel scheduler for a general-purpose OS is no longer the root scheduler; this privilege is taken over by a small hard-real-time OS. …”

“…Very low dispatch latency can be achieved using co-resident operating systems [11,93]. This approach virtualizes the interrupt controller seen by a general-purpose operating system in order to allow a small real-time kernel to run even when the GPOS has "disabled interrupts."…”

Intelligent OS Process Scheduling Using Fuzzy Inference with User Models

“…SMART proposed by Nieh and Monica is a hierarchical CPU scheduling algorithm that supports both hard real time and conventional time sharing applications, adjusts well to overload, and can notify applications when their deadlines cannot be met [7]. Other scheduling work has been done in supporting clients with real-time requirements [8,9] and improving the response time of interactive clients [10,11]. …”

The performance trade-offs of implementing a large scale real-time application using the Windows NT operating system

“…One where both the real-time and non-realtime processing co-exist on the same platform [13] [12] [14] [15] [16]. The non-real-time processing uses the Windows NT operating system, and all real-time tasks run under a real-time extension to Windows NT.…”