Deadline scheduling in fault tolerant real time systems
Klaus Gresser, Hella Thielen
Abstract
Klaus Gresser, Hella Thielen
Abstract
Deadline scheduling means that the task with the earliest required response time is processed (earliest deadline first). In many points this kind of scheduling is superior to fixed priority scheduling. This work investigates the applicability of dynamic deadline scheduling in fault tolerant multicomputer systems with active redundancy and asynchronous scheduling. Tasks are preemptive and are subject to precedence constraints. It is shown that the additional computation resources necessary for fault situations can be reduced by relating the error detection timer to the deadline. If however the usual methods of deadline scheduling are applied deadlines can be missed. A new technique of deadline scheduling meeting all requirements for fault tolerant systems is proposed.
OpenAlex reports 4 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
Deadline scheduling means that the task with the earliest required response time is processed (earliest deadline first). In many points this kind of scheduling is superior to fixed priority scheduling. This work investigates the applicability of dynamic deadline scheduling in fault tolerant multicomputer systems with active redundancy and asynchronous scheduling. Tasks are preemptive and are subject to precedence constraints. It is shown that the additional computation resources necessary for fault situations can be reduced by relating the error detection timer to the deadline. If however the usual methods of deadline scheduling are applied deadlines can be missed. A new technique of deadline scheduling meeting all requirements for fault tolerant systems is proposed.
Key concepts: Computer science, Earliest deadline first scheduling, Fixed-priority pre-emptive scheduling, Dynamic priority scheduling, Deadline-monotonic scheduling, Rate-monotonic scheduling, Fair-share scheduling, Distributed computing