EDZL scheduling analysis

Abstract

A schedulability test is derived for the global Earliest Deadline Zero Laxity (EDZL) scheduling algorithm on a platform with multiple identical processors. The test is sufficient, but not necessary, to guarantee that a system of independent sporadic tasks with arbitrary deadlines will be successfully scheduled, with no missed deadlines, by the multiprocessor EDZL algorithm. Global EDZL is known to be at least as effective as global Earliest-Deadline-First (EDF) in scheduling task sets to meet deadlines. It is shown, by testing on large numbers of pseudo-randomly generated task sets, that the combination of EDZL and the new schedulability test is able to guarantee that far more task sets meet deadlines than the combination of EDF and known EDF schedulability tests.In the second part of the paper, an improved version of the EDZL-schedulability test is presented. This new algorithm is able to efficiently exploit information on the slack values of interfering tasks, to iteratively refine the estimation of the interference a task can be subjected to. This iterative algorithm is shown to have better performance than the initial test, in terms of schedulable task sets detected. This is an extended version of the ECRTS'07 paper of Cirinei and Baker (2007), with corrections of some flaws in that original paper and a new iterative schedulability test based on Cirinei (2007).

FPSL, FPCL and FPZL schedulability analysis

“…This test reduces the over-estimation of carryin interference, a feature of the previous tests, by iteratively calculating a lower bound on the slack for each task. The empirical evaluation by Baker et al (2008) shows that this iterative test for EDZL outperforms other tests for EDZL (Cirinei and Baker, 2007) and as expected, similar tests for global EDF. Kato and Yamasaki (2008), introduced EDCL; a variant of EDZL, which increases job priority on the basis of laxity at the release or completion time of a job.…”

“…The Optimality Degree of each algorithm / schedulability test is given in Table 1 below, expressed as a percentage. Table 1 shows that the Optimality Degree for FPZL scheduling using the polynomial time DA-LC schedulability test derived in this paper, (with OPA priority assignment and zero-laxity execution time calculation) is 3-4% better than for global FP scheduling using OPA and an equivalent schedulability test, and 13% better than for EDZL, assuming the iterative schedulability test given by Baker et al (2008). By comparison, FPSL scheduling has an Optimality Degree that is approx.…”

“…Piao et al (2006) showed that EDZL is also completion time predictable A simpler proof of predictability was given by Cirinei and Baker (2007), who also developed a sufficient schedulability test for EDZL based on the fundamental strategy of Baker (2003). Baker et al (2008) gave an iterative sufficient test for EDZL based on the approach taken by Bertogna (2007) and Bertogna et al (2009) for work conserving algorithms and global EDF. This test reduces the over-estimation of carryin interference, a feature of the previous tests, by iteratively calculating a lower bound on the slack for each task.…”

“…We investigated the performance of the FPSL and FPZL DA-LC, schedulability tests using the OPA algorithm (Audsley, 1991(Audsley, , 2001) to assign priorities, and compared their performance to that of the equivalent test for global FP scheduling, and to schedulability tests for global EDF by Bertogna et al (2009) (the "EDF-RTA" test) and EDZL scheduling Baker et al (2008) (the "EDZL-I test"). Also shown on the graphs are results for the necessary infeasibility test of Baker and Cirinei (2006) (labelled "LOAD*").…”

Laxity dynamics and LLF schedulability analysis on multiprocessor platforms

“…2. We characterize these properties using conditions on the worst-case higherpriority interference, because previous studies suggest that it is feasible to derive multiprocessor schedulability tests using such conditions (e.g., Baker et al 2008). …”