Efficient alternatives for implementing fixed-priority schedulers

  • Authors:
  • Sergio Sáez;Vicent Lorente;Silvia Terrasa;Alfons Crespo

  • Affiliations:
  • Universidad Politécnica de Valencia, Valencia, Spain;Universidad Politécnica de Valencia, Valencia, Spain;Universidad Politécnica de Valencia, Valencia, Spain;Universidad Politécnica de Valencia, Valencia, Spain

  • Venue:
  • Ada-Europe'05 Proceedings of the 10th Ada-Europe international conference on Reliable Software Technologies
  • Year:
  • 2005

Quantified Score

Hi-index 0.00

Visualization

Abstract

Fixed-priority schedulers received a lot of attention from the real-time community, and a big effort has been performed to develop accurate and more general schedulability analysis that can ensure the correct execution of the system. Nevertheless, only few works analysed the overhead and blocking intervals introduced by the scheduler and the associated kernel routines, modifying the schedulability tests to take into account these overheads. However, all these works assume a very simple scheduler that uses trivial data structures to stores the tasks information. This work analyses data structures used in several open real-time kernels that supports Ada language. Additionally, a new data structure, the Cartesian trees, is proposed. As a conclusion, the preliminary studies show that alternative data structures, as proposed Cartesian trees, can improve the scalability of the system strongly reducing the scheduling overheads.