Prediction-Based Power Oversubscription in Cloud Platforms

Datacenter designers rely on conservative estimates of IT equipment power draw to provision resources. This leaves resources underutilized and requires more datacenters to be built. Prior work has used power capping to shave the rare power peaks and add more servers to the datacenter, thereby oversu...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Kumbhare, Alok, Azimi, Reza, Manousakis, Ioannis, Bonde, Anand, Frujeri, Felipe, Mahalingam, Nithish, Misra, Pulkit, Javadi, Seyyed Ahmad, Schroeder, Bianca, Fontoura, Marcus, Bianchini, Ricardo
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page
container_issue
container_start_page
container_title
container_volume
creator Kumbhare, Alok
Azimi, Reza
Manousakis, Ioannis
Bonde, Anand
Frujeri, Felipe
Mahalingam, Nithish
Misra, Pulkit
Javadi, Seyyed Ahmad
Schroeder, Bianca
Fontoura, Marcus
Bianchini, Ricardo
description Datacenter designers rely on conservative estimates of IT equipment power draw to provision resources. This leaves resources underutilized and requires more datacenters to be built. Prior work has used power capping to shave the rare power peaks and add more servers to the datacenter, thereby oversubscribing its resources and lowering capital costs. This works well when the workloads and their server placements are known. Unfortunately, these factors are unknown in public clouds, forcing providers to limit the oversubscription so that performance is never impacted. In this paper, we argue that providers can use predictions of workload performance criticality and virtual machine (VM) resource utilization to increase oversubscription. This poses many challenges, such as identifying the performance-critical workloads from black-box VMs, creating support for criticality-aware power management, and increasing oversubscription while limiting the impact of capping. We address these challenges for the hardware and software infrastructures of Microsoft Azure. The results show that we enable a 2x increase in oversubscription with minimum impact to critical workloads.
doi_str_mv 10.48550/arxiv.2010.15388
format Article
fullrecord <record><control><sourceid>arxiv_GOX</sourceid><recordid>TN_cdi_arxiv_primary_2010_15388</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2010_15388</sourcerecordid><originalsourceid>FETCH-LOGICAL-a678-3c8ec86ebe69ba5c41179fb13e0c861a34d7b85f7e9a19f5376c6d94b91a2a663</originalsourceid><addsrcrecordid>eNotj8kKwjAURbNxIeoHuLI_UG1MM7nT4gSCLtyXl_QFAtVK4vj31mF14Vy43EPIkGbjXHGeTSA8_X08zVpAOVOqS2aHgJW3V9-c0wVErJJD88CQ7O8Y4s1EG_zlUyb-nBR1c2v7Gq6uCafYJx0HdcTBP3vkuFoei02626-3xXyXgpAqZVahVQINCm2A25xSqZ2hDLMWU2B5JY3iTqIGqh1nUlhR6dxoClMQgvXI6Df7PV9egj9BeJUfifIrwd6BVUIY</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>Prediction-Based Power Oversubscription in Cloud Platforms</title><source>arXiv.org</source><creator>Kumbhare, Alok ; Azimi, Reza ; Manousakis, Ioannis ; Bonde, Anand ; Frujeri, Felipe ; Mahalingam, Nithish ; Misra, Pulkit ; Javadi, Seyyed Ahmad ; Schroeder, Bianca ; Fontoura, Marcus ; Bianchini, Ricardo</creator><creatorcontrib>Kumbhare, Alok ; Azimi, Reza ; Manousakis, Ioannis ; Bonde, Anand ; Frujeri, Felipe ; Mahalingam, Nithish ; Misra, Pulkit ; Javadi, Seyyed Ahmad ; Schroeder, Bianca ; Fontoura, Marcus ; Bianchini, Ricardo</creatorcontrib><description>Datacenter designers rely on conservative estimates of IT equipment power draw to provision resources. This leaves resources underutilized and requires more datacenters to be built. Prior work has used power capping to shave the rare power peaks and add more servers to the datacenter, thereby oversubscribing its resources and lowering capital costs. This works well when the workloads and their server placements are known. Unfortunately, these factors are unknown in public clouds, forcing providers to limit the oversubscription so that performance is never impacted. In this paper, we argue that providers can use predictions of workload performance criticality and virtual machine (VM) resource utilization to increase oversubscription. This poses many challenges, such as identifying the performance-critical workloads from black-box VMs, creating support for criticality-aware power management, and increasing oversubscription while limiting the impact of capping. We address these challenges for the hardware and software infrastructures of Microsoft Azure. The results show that we enable a 2x increase in oversubscription with minimum impact to critical workloads.</description><identifier>DOI: 10.48550/arxiv.2010.15388</identifier><language>eng</language><subject>Computer Science - Distributed, Parallel, and Cluster Computing</subject><creationdate>2020-10</creationdate><rights>http://arxiv.org/licenses/nonexclusive-distrib/1.0</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,777,882</link.rule.ids><linktorsrc>$$Uhttps://arxiv.org/abs/2010.15388$$EView_record_in_Cornell_University$$FView_record_in_$$GCornell_University$$Hfree_for_read</linktorsrc><backlink>$$Uhttps://doi.org/10.48550/arXiv.2010.15388$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Kumbhare, Alok</creatorcontrib><creatorcontrib>Azimi, Reza</creatorcontrib><creatorcontrib>Manousakis, Ioannis</creatorcontrib><creatorcontrib>Bonde, Anand</creatorcontrib><creatorcontrib>Frujeri, Felipe</creatorcontrib><creatorcontrib>Mahalingam, Nithish</creatorcontrib><creatorcontrib>Misra, Pulkit</creatorcontrib><creatorcontrib>Javadi, Seyyed Ahmad</creatorcontrib><creatorcontrib>Schroeder, Bianca</creatorcontrib><creatorcontrib>Fontoura, Marcus</creatorcontrib><creatorcontrib>Bianchini, Ricardo</creatorcontrib><title>Prediction-Based Power Oversubscription in Cloud Platforms</title><description>Datacenter designers rely on conservative estimates of IT equipment power draw to provision resources. This leaves resources underutilized and requires more datacenters to be built. Prior work has used power capping to shave the rare power peaks and add more servers to the datacenter, thereby oversubscribing its resources and lowering capital costs. This works well when the workloads and their server placements are known. Unfortunately, these factors are unknown in public clouds, forcing providers to limit the oversubscription so that performance is never impacted. In this paper, we argue that providers can use predictions of workload performance criticality and virtual machine (VM) resource utilization to increase oversubscription. This poses many challenges, such as identifying the performance-critical workloads from black-box VMs, creating support for criticality-aware power management, and increasing oversubscription while limiting the impact of capping. We address these challenges for the hardware and software infrastructures of Microsoft Azure. The results show that we enable a 2x increase in oversubscription with minimum impact to critical workloads.</description><subject>Computer Science - Distributed, Parallel, and Cluster Computing</subject><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2020</creationdate><recordtype>article</recordtype><sourceid>GOX</sourceid><recordid>eNotj8kKwjAURbNxIeoHuLI_UG1MM7nT4gSCLtyXl_QFAtVK4vj31mF14Vy43EPIkGbjXHGeTSA8_X08zVpAOVOqS2aHgJW3V9-c0wVErJJD88CQ7O8Y4s1EG_zlUyb-nBR1c2v7Gq6uCafYJx0HdcTBP3vkuFoei02626-3xXyXgpAqZVahVQINCm2A25xSqZ2hDLMWU2B5JY3iTqIGqh1nUlhR6dxoClMQgvXI6Df7PV9egj9BeJUfifIrwd6BVUIY</recordid><startdate>20201029</startdate><enddate>20201029</enddate><creator>Kumbhare, Alok</creator><creator>Azimi, Reza</creator><creator>Manousakis, Ioannis</creator><creator>Bonde, Anand</creator><creator>Frujeri, Felipe</creator><creator>Mahalingam, Nithish</creator><creator>Misra, Pulkit</creator><creator>Javadi, Seyyed Ahmad</creator><creator>Schroeder, Bianca</creator><creator>Fontoura, Marcus</creator><creator>Bianchini, Ricardo</creator><scope>AKY</scope><scope>GOX</scope></search><sort><creationdate>20201029</creationdate><title>Prediction-Based Power Oversubscription in Cloud Platforms</title><author>Kumbhare, Alok ; Azimi, Reza ; Manousakis, Ioannis ; Bonde, Anand ; Frujeri, Felipe ; Mahalingam, Nithish ; Misra, Pulkit ; Javadi, Seyyed Ahmad ; Schroeder, Bianca ; Fontoura, Marcus ; Bianchini, Ricardo</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-a678-3c8ec86ebe69ba5c41179fb13e0c861a34d7b85f7e9a19f5376c6d94b91a2a663</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2020</creationdate><topic>Computer Science - Distributed, Parallel, and Cluster Computing</topic><toplevel>online_resources</toplevel><creatorcontrib>Kumbhare, Alok</creatorcontrib><creatorcontrib>Azimi, Reza</creatorcontrib><creatorcontrib>Manousakis, Ioannis</creatorcontrib><creatorcontrib>Bonde, Anand</creatorcontrib><creatorcontrib>Frujeri, Felipe</creatorcontrib><creatorcontrib>Mahalingam, Nithish</creatorcontrib><creatorcontrib>Misra, Pulkit</creatorcontrib><creatorcontrib>Javadi, Seyyed Ahmad</creatorcontrib><creatorcontrib>Schroeder, Bianca</creatorcontrib><creatorcontrib>Fontoura, Marcus</creatorcontrib><creatorcontrib>Bianchini, Ricardo</creatorcontrib><collection>arXiv Computer Science</collection><collection>arXiv.org</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Kumbhare, Alok</au><au>Azimi, Reza</au><au>Manousakis, Ioannis</au><au>Bonde, Anand</au><au>Frujeri, Felipe</au><au>Mahalingam, Nithish</au><au>Misra, Pulkit</au><au>Javadi, Seyyed Ahmad</au><au>Schroeder, Bianca</au><au>Fontoura, Marcus</au><au>Bianchini, Ricardo</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Prediction-Based Power Oversubscription in Cloud Platforms</atitle><date>2020-10-29</date><risdate>2020</risdate><abstract>Datacenter designers rely on conservative estimates of IT equipment power draw to provision resources. This leaves resources underutilized and requires more datacenters to be built. Prior work has used power capping to shave the rare power peaks and add more servers to the datacenter, thereby oversubscribing its resources and lowering capital costs. This works well when the workloads and their server placements are known. Unfortunately, these factors are unknown in public clouds, forcing providers to limit the oversubscription so that performance is never impacted. In this paper, we argue that providers can use predictions of workload performance criticality and virtual machine (VM) resource utilization to increase oversubscription. This poses many challenges, such as identifying the performance-critical workloads from black-box VMs, creating support for criticality-aware power management, and increasing oversubscription while limiting the impact of capping. We address these challenges for the hardware and software infrastructures of Microsoft Azure. The results show that we enable a 2x increase in oversubscription with minimum impact to critical workloads.</abstract><doi>10.48550/arxiv.2010.15388</doi><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier DOI: 10.48550/arxiv.2010.15388
ispartof
issn
language eng
recordid cdi_arxiv_primary_2010_15388
source arXiv.org
subjects Computer Science - Distributed, Parallel, and Cluster Computing
title Prediction-Based Power Oversubscription in Cloud Platforms
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-20T20%3A58%3A44IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-arxiv_GOX&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Prediction-Based%20Power%20Oversubscription%20in%20Cloud%20Platforms&rft.au=Kumbhare,%20Alok&rft.date=2020-10-29&rft_id=info:doi/10.48550/arxiv.2010.15388&rft_dat=%3Carxiv_GOX%3E2010_15388%3C/arxiv_GOX%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true