Real-time atomic multicast algorithms implemented on a shared memory multiprocessor

P.D.V. Stok, van der, A.H.L. Janssen-Raemaekers

Research output: Contribution to journalArticleAcademicpeer-review

Abstract

Processes are distributed over processors that share directly accessible memory. Processes exchange messages via memory. Algorithms are specified which satisfy the atomic multicast requirements under two fault hypotheses: (1) fail-stop memory and fail stop processes and (2) fail-stop memory and processes with timing failures. It is shown that the algorithms can be applied under different scheduling conditions: (1) hard real-time (HRT) applications composed of periodically scheduled tasks and (2) HRT applications with tasks scheduled according to a static off-line calculated schedule. The performance measurements on an implementation of two algorithms are presented.
Original languageEnglish
Pages (from-to)55-91
JournalReal-Time Systems
Volume24
Issue number1
DOIs
Publication statusPublished - 2003

Fingerprint Dive into the research topics of 'Real-time atomic multicast algorithms implemented on a shared memory multiprocessor'. Together they form a unique fingerprint.

  • Cite this