• DocumentCode
    3173348
  • Title

    Resource management for enhancing predictability in systems with limited processing capabilities

  • Author

    Alonso, Alejandro ; Salazar, Emilio ; López, Jorge

  • Author_Institution
    Dept. Ing. de Sist. Telematicos, Univ. Politec. de Madrid, Madrid, Spain
  • fYear
    2010
  • fDate
    13-16 Sept. 2010
  • Firstpage
    1
  • Lastpage
    7
  • Abstract
    There is an increasing demand for computing systems composed by heterogeneous computers, connected by different types of networks, and that allow for accessing a wide range of services in a seamless way. Some of those computers are mobile or embedded and have limited resources, and can be overloaded when trying to handle their users demands. Then it is not possible to ensure a proper behaviour of the running applications. This can be an important problem when dealing with critical events in healthcare, home surveillance, or forest monitoring. Resource reservation is a valid basis for handling this issue. It allows for guaranteeing a certain resource share for applications that are important for the proper behavior of a given system. This paper describes an implementation of a resource management component and its integration in the Linux kernel. This piece of software has allowed to assign CPU budgets to standard Java threads, which is an important facility, given the widespread of this programming language. This implementation has been validated on service oriented middleware, where relevant services are executed by thread with guaranteed budget, to improve its predictability.
  • Keywords
    Java; Linux; middleware; resource allocation; CPU budgets; Java; Linux kernel; computing systems; heterogeneous computers; resource management; resource reservation; service oriented middleware; system predictability;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Emerging Technologies and Factory Automation (ETFA), 2010 IEEE Conference on
  • Conference_Location
    Bilbao
  • ISSN
    1946-0740
  • Print_ISBN
    978-1-4244-6848-5
  • Type

    conf

  • DOI
    10.1109/ETFA.2010.5641339
  • Filename
    5641339