A resource management framework for priority-based physical-memory allocation

Authors

Kingsley Cheung and Gernot Heiser

    School of Computer Science and Engineering
    UNSW,
    Sydney 2052, Australia

Abstract

Most multitasking operating systems support scheduling priorities in order to ensure that processor time is allocated to important or time-critical processes in preference to less important ones. Ideally this would prevent a low-priority process from slowing the execution of a high-priority one. In practice, strict prioritisation is undermined by a lack of suitable allocation policy for resources other than CPU time. For example, a low priority process may degrade the execution speed of a high-priority process by competing with it for physical memory. We present the design of a flexible resource management framework which prioritises memory allocation, and examine a prototype implementation for the Mungi single-address-space operating system.

BibTeX Entry

  @inproceedings{Cheung_Heiser_02,
    author           = {Kingsley Cheung and Gernot Heiser},
    title            = {A Resource Management Framework for Priority-Based Physical-Memory Allocation},
    month            = {jan},
    editor           = {J Morris and F Lai},
    year             = {2002},
    booktitle        = {Proceedings of the 7th Asia-Pacific Computer Systems Architecture Conference},
    address          = {Monash University, Melbourne, Australia}
  }

Download

Served by Apache on Linux on seL4