Batching earliest deadline first scheduling

Abstract

Investigates the trade-off in the dynamic scheduling of real-time tasks, between the frequency at which the scheduling algorithm is invoked, the size of the task set to which the scheduling (prioritization) policy is applied at every invocation, and the quality of the resulting schedules in terms of deadline compliance. We identify two classes of algorithms, one of which forms a batch of arrived tasks and which schedules and executes all tasks in a batch before considering other tasks that arrive in the meantime. The other class accounts for and schedules arrived tasks more frequently and applies the scheduling policy to all available tasks. We compare the performance of a batching and a non-batching technique, both of which apply an earliest-deadline-first (EDF) policy to prioritize tasks. An experimental evaluation of the proposed algorithms shows that our batching algorithms outperform their non-batching counterparts under tighter time constraints.