An operating system (OS) is a supervisory program in a computing system, responsible for efficient management of the hardware resources. In the context of real-time systems, that is, systems in which timeliness and predictability in the worst case are critical, the real-time OS (RTOS) additionally has to ensure satisfaction of all hard deadlines in the system. This chapter considers the RTOS resource scheduling problem in a variety of computing architectures including single-core central processing units (CPUs), multi-core CPUs, CPUs with graphics processing units (GPUs) as co-processors and distributed edge servers. In particular, seminal literature addressing the problem of real-time scheduling of the processing capacity in these architectures is presented. A review of important resource management techniques for wireless and wired networks with real-time requirements is also presented, since such networks are essential for the predictable transmission of workload in the distributed edge.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Real-Time Scheduling for Computing Architectures

  • Arvind Easwaran,
  • Michael Yuhas,
  • Saravanan Ramanathan,
  • Ankita Samaddar

摘要

An operating system (OS) is a supervisory program in a computing system, responsible for efficient management of the hardware resources. In the context of real-time systems, that is, systems in which timeliness and predictability in the worst case are critical, the real-time OS (RTOS) additionally has to ensure satisfaction of all hard deadlines in the system. This chapter considers the RTOS resource scheduling problem in a variety of computing architectures including single-core central processing units (CPUs), multi-core CPUs, CPUs with graphics processing units (GPUs) as co-processors and distributed edge servers. In particular, seminal literature addressing the problem of real-time scheduling of the processing capacity in these architectures is presented. A review of important resource management techniques for wireless and wired networks with real-time requirements is also presented, since such networks are essential for the predictable transmission of workload in the distributed edge.