Abstract:
This paper focuses on reducing the computational cost of iterative algorithms used to evaluate the schedulability of a Real-Time System scheduled by the Rate Monotonic (R...Show MoreMetadata
Abstract:
This paper focuses on reducing the computational cost of iterative algorithms used to evaluate the schedulability of a Real-Time System scheduled by the Rate Monotonic (RM) or Deadline Monotonic (DM) policies. These algorithms calculate the worst case response time of each task of a real-time system. Several runtime strategies for fault tolerance mechanisms, power consumption methods and execution of non-real-time tasks, among others, require low computational cost schedulability algorithms to be efficiently implemented in real applications. The algorithm proposed is based in mathematical properties of the real-time schedulability test, that reduce noticeably the number of iterations needed to converge to the final solution. Through simulations, it is found that the proposed algorithm produces a significant reduction in the average temporal cost reaching in some cases, a complexity reduction from O(n2) to O(n.log(n)), with respect to classical schedulability algorithms.
Date of Conference: 09-11 August 2017
Date Added to IEEE Xplore: 23 November 2017
ISBN Information: