Distributed resource management across process boundaries

Handle URI:
http://hdl.handle.net/10754/625905
Title:
Distributed resource management across process boundaries
Authors:
Suresh, Lalith; Bodik, Peter; Menache, Ishai; Canini, Marco ( 0000-0002-5051-4283 ) ; Ciucu, Florin
Abstract:
Multi-tenant distributed systems composed of small services, such as Service-oriented Architectures (SOAs) and Micro-services, raise new challenges in attaining high performance and efficient resource utilization. In these systems, a request execution spans tens to thousands of processes, and the execution paths and resource demands on different services are generally not known when a request first enters the system. In this paper, we highlight the fundamental challenges of regulating load and scheduling in SOAs while meeting end-to-end performance objectives on metrics of concern to both tenants and operators. We design Wisp, a framework for building SOAs that transparently adapts rate limiters and request schedulers system-wide according to operator policies to satisfy end-to-end goals while responding to changing system conditions. In evaluations against production as well as synthetic workloads, Wisp successfully enforces a range of end-to-end performance objectives, such as reducing average latencies, meeting deadlines, providing fairness and isolation, and avoiding system overload.
KAUST Department:
KAUST
Citation:
Suresh L, Bodik P, Menache I, Canini M, Ciucu F (2017) Distributed resource management across process boundaries. Proceedings of the 2017 Symposium on Cloud Computing - SoCC ’17. Available: http://dx.doi.org/10.1145/3127479.3132020.
Publisher:
ACM Press
Journal:
Proceedings of the 2017 Symposium on Cloud Computing - SoCC '17
Issue Date:
27-Sep-2017
DOI:
10.1145/3127479.3132020
Type:
Conference Paper
Sponsors:
We thank our shepherd, Iqbal Mohomed, and the anonymous reviewers for their feedback. We thank Madan Musuvathi, Srikanth Kandula, and Virajith Jalaparti for the useful discussions that helped shape this paper.
Additional Links:
https://dl.acm.org/citation.cfm?doid=3127479.3132020
Appears in Collections:
Conference Papers

Full metadata record

DC FieldValue Language
dc.contributor.authorSuresh, Lalithen
dc.contributor.authorBodik, Peteren
dc.contributor.authorMenache, Ishaien
dc.contributor.authorCanini, Marcoen
dc.contributor.authorCiucu, Florinen
dc.date.accessioned2017-10-19T07:10:41Z-
dc.date.available2017-10-19T07:10:41Z-
dc.date.issued2017-09-27en
dc.identifier.citationSuresh L, Bodik P, Menache I, Canini M, Ciucu F (2017) Distributed resource management across process boundaries. Proceedings of the 2017 Symposium on Cloud Computing - SoCC ’17. Available: http://dx.doi.org/10.1145/3127479.3132020.en
dc.identifier.doi10.1145/3127479.3132020en
dc.identifier.urihttp://hdl.handle.net/10754/625905-
dc.description.abstractMulti-tenant distributed systems composed of small services, such as Service-oriented Architectures (SOAs) and Micro-services, raise new challenges in attaining high performance and efficient resource utilization. In these systems, a request execution spans tens to thousands of processes, and the execution paths and resource demands on different services are generally not known when a request first enters the system. In this paper, we highlight the fundamental challenges of regulating load and scheduling in SOAs while meeting end-to-end performance objectives on metrics of concern to both tenants and operators. We design Wisp, a framework for building SOAs that transparently adapts rate limiters and request schedulers system-wide according to operator policies to satisfy end-to-end goals while responding to changing system conditions. In evaluations against production as well as synthetic workloads, Wisp successfully enforces a range of end-to-end performance objectives, such as reducing average latencies, meeting deadlines, providing fairness and isolation, and avoiding system overload.en
dc.description.sponsorshipWe thank our shepherd, Iqbal Mohomed, and the anonymous reviewers for their feedback. We thank Madan Musuvathi, Srikanth Kandula, and Virajith Jalaparti for the useful discussions that helped shape this paper.en
dc.publisherACM Pressen
dc.relation.urlhttps://dl.acm.org/citation.cfm?doid=3127479.3132020en
dc.rightsArchived with thanks to Proceedings of the 2017 Symposium on Cloud Computing - SoCC '17en
dc.subjectMicroservicesen
dc.subjectService-Oriented Architecturesen
dc.subjectResource Managementen
dc.subjectRate Limitingen
dc.subjectSchedulingen
dc.titleDistributed resource management across process boundariesen
dc.typeConference Paperen
dc.contributor.departmentKAUSTen
dc.identifier.journalProceedings of the 2017 Symposium on Cloud Computing - SoCC '17en
dc.eprint.versionPublisher's Version/PDFen
dc.contributor.institutionMicrosoft Researchen
dc.contributor.institutionUniversity of Warwicken
kaust.authorCanini, Marcoen
All Items in KAUST are protected by copyright, with all rights reserved, unless otherwise indicated.