Performance Evaluation of the Weakly Hard Real-Time Tasks for Global Multiprocessor Scheduling Approach

Real-time systems can be classified into three categories, based on the “seriousness” of deadline misses either by hard, soft, or weakly hard real-time tasks. The consequences of deadlines miss for a hard real-time task cannot be tolerated because some failure can affect the system performance whereas some deadline misses can be tolerated for soft real-time tasks. Meanwhile, in a weakly hard real-time task, the distribution of its met and missed deadlines is stated and specified precisely. Due to the complexity and significantly increased functionality in system computation, attention has been given to multiprocessor scheduling. Studies have shown that current multiprocessor scheduling of weakly hard real-time tasks used imprecise computation model based on iterative algorithms. This algorithm decomposed into two parts; mandatory and optional parts, unfortunately, the result analysis is precise only if its mandatory and optional parts are both executed. Even, the use of hierarchical scheduling algorithm, such as two-level scheduling under PFair algorithm may cause high overhead due to frequent preemptions and migrations. Furthermore, this algorithm incurs significant run-time overhead due to their quantum-based scheduling. In order to cater for the limitations and stated problems, an alternative multiprocessor scheduling approach, called global scheduling is proposed. The proposed scheduling approach aims to improve the probability of deadline satisfactions as much as possible and at the same time achieve a higher utilization of the task sets, with less task migrations. Thus, in this paper, performance measurement parameters are used as performance evaluation of the proposed scheduling approach.