Ítem
Solo Metadatos

Differential approximation and sprinting for multi-priority big data engines

Mostrar el registro sencillo de la publicación
dc.creatorBirke R.spa
dc.creatorRocha I.spa
dc.creatorPérez, Juan F.spa
dc.creatorSchiavoni V.spa
dc.creatorFelber P.spa
dc.creatorChen L.Y.spa
dc.date.accessioned2020-05-25T23:57:04Z
dc.date.available2020-05-25T23:57:04Z
dc.date.created2019spa
dc.description.abstractToday’s big data clusters based on the MapReduce paradigm are capable of executing analysis jobs with multiple priorities, providing differential latency guarantees. Traces from production systems show that the latency advantage of high-priority jobs comes at the cost of severe latency degradation of low-priority jobs as well as daunting resource waste caused by repetitive eviction and re-execution of low-priority jobs. We advocate a new resource management design that exploits the idea of differential approximation and sprinting. The unique combination of approximation and sprinting avoids the eviction of low-priority jobs and its consequent latency degradation and resource waste. To this end, we designed, implemented and evaluated DiAS, an extension of the Spark processing engine to support deflate jobs by dropping tasks and to sprint jobs. Our experiments on scenarios with two and three priority classes indicate that DiAS achieves up to 90% and 60% latency reduction for low- and high-priority jobs, respectively. DiAS not only eliminates resource waste but also (surprisingly) lowers energy consumption up to 30% at only a marginal accuracy loss for low-priority jobs. © 2019 Association for Computing Machinery.eng
dc.format.mimetypeapplication/pdf
dc.identifier.doihttps://doi.org/10.1145/3361525.3361547
dc.identifier.urihttps://repository.urosario.edu.co/handle/10336/22595
dc.language.isoengspa
dc.publisherAssociation for Computing Machinery, Incspa
dc.relation.citationEndPage214
dc.relation.citationStartPage202
dc.relation.citationTitleMiddleware 2019 - Proceedings of the 2019 20th International Middleware Conference
dc.relation.ispartofMiddleware 2019 - Proceedings of the 2019 20th International Middleware Conference,(2019); pp. 202-214spa
dc.relation.urihttps://www.scopus.com/inward/record.uri?eid=2-s2.0-85078012099&doi=10.1145%2f3361525.3361547&partnerID=40&md5=8daf0251625a159a239e3b82ad74d3a6spa
dc.rights.accesRightsinfo:eu-repo/semantics/openAccess
dc.rights.accesoAbierto (Texto Completo)spa
dc.source.instnameinstname:Universidad del Rosariospa
dc.source.reponamereponame:Repositorio Institucional EdocURspa
dc.subject.keywordElectric sparksspa
dc.subject.keywordEnergy conservationspa
dc.subject.keywordEnergy utilizationspa
dc.subject.keywordEnginesspa
dc.subject.keywordMiddlewarespa
dc.subject.keywordDifferential approximationsspa
dc.subject.keywordLatency reductionspa
dc.subject.keywordPrioritiesspa
dc.subject.keywordProduction systemspa
dc.subject.keywordResource managementspa
dc.subject.keywordResource wastesspa
dc.subject.keywordSpark processingspa
dc.subject.keywordSprintingspa
dc.subject.keywordBig dataspa
dc.subject.keywordDifferential approximationspa
dc.subject.keywordEnergy savingsspa
dc.subject.keywordPrioritiesspa
dc.subject.keywordSparkspa
dc.subject.keywordSprintingspa
dc.titleDifferential approximation and sprinting for multi-priority big data enginesspa
dc.typeconferenceObjecteng
dc.type.hasVersioninfo:eu-repo/semantics/publishedVersion
dc.type.spaDocumento de conferenciaspa
Archivos
Colecciones
Artículos