Barely alive memory servers

Author:

Anagnostopoulou Vlasia1,Biswas Susmit1,Saadeldeen Heba1,Savage Alan1,Bianchini Ricardo2,Yang Tao1,Franklin Diana1,Chong Frederic T.1

Affiliation:

1. University of California at Santa Barbara

2. Rutgers University

Abstract

Current resource provisioning schemes in Internet services leave servers less than 50% utilized almost all the time. At this level of utilization, the servers' energy efficiency is substantially lower than at peak utilization. A solution to this problem could be dynamically consolidating workloads into fewer servers and turning others off. However, services typically resist doing so, because of high response times during reactivation in handling traffic spikes. Moreover, services often want the memory and/or storage of all servers to be readily available at all times. In this article, we propose a family of barely alive active low-power server states that facilitates both fast reactivation and access to memory while in a low-power state. We compare these states to previously proposed active and idle states. In particular, we investigate the impact of load bursts in each energy-saving scheme. We also evaluate the additional benefits of memory access under low-power states with a study of a search service using a cooperative main-memory cache. Finally, we propose a system that combines a barely-alive state with the off state. We find that the barely alive states can reduce service energy consumption by up to 38%, compared to an energy-oblivious system. We also find that these energy savings are consistent across a large parameter space.

Funder

Energy-Proportional Computing

Publisher

Association for Computing Machinery (ACM)

Subject

Electrical and Electronic Engineering,Hardware and Architecture,Software

Reference31 articles.

1. Adamic L. 2000. Zipf power-laws and pareto -- A ranking tutorial. Tech. rep. HP Labs. Adamic L. 2000. Zipf power-laws and pareto -- A ranking tutorial. Tech. rep. HP Labs.

2. FAWN

3. The Case for Energy-Proportional Computing

4. PRESS: a clustered server based on user-level communication

Cited by 12 articles. 订阅此论文施引文献 订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献

1. Pelican: Power Scheduling for QoS in Large-scale Data Centers with Heterogeneous Workloads;2019 Tenth International Green and Sustainable Computing Conference (IGSC);2019-10

2. What Your DRAM Power Models Are Not Telling You;Proceedings of the ACM on Measurement and Analysis of Computing Systems;2018-12-21

3. Welcome to zombieland;Proceedings of the Thirteenth EuroSys Conference;2018-04-23

4. Peak efficiency aware scheduling for highly energy proportional servers;ACM SIGARCH Computer Architecture News;2016-10-12

5. On Energy Proportionality and Time-Energy Performance of Heterogeneous Clusters;2016 IEEE International Conference on Cluster Computing (CLUSTER);2016-09

同舟云学术

1.学者识别学者识别

2.学术分析学术分析

3.人才评估人才评估

"同舟云学术"是以全球学者为主线,采集、加工和组织学术论文而形成的新型学术文献查询和分析系统,可以对全球学者进行文献检索和人才价值评估。用户可以通过关注某些学科领域的顶尖人物而持续追踪该领域的学科进展和研究前沿。经过近期的数据扩容,当前同舟云学术共收录了国内外主流学术期刊6万余种,收集的期刊论文及会议论文总量共计约1.5亿篇,并以每天添加12000余篇中外论文的速度递增。我们也可以为用户提供个性化、定制化的学者数据。欢迎来电咨询!咨询电话:010-8811{复制后删除}0370

www.globalauthorid.com

TOP

Copyright © 2019-2024 北京同舟云网络信息技术有限公司
京公网安备11010802033243号  京ICP备18003416号-3