基于PE方法的数据中心需量费用优化算法

doi:10.11959/j.issn.1000-436x.2016056

通信学报 ›› 2016, Vol. 37 ›› Issue (3): 90-97.doi: 10.11959/j.issn.1000-436x.2016056

基于PE方法的数据中心需量费用优化算法

黄焱^1,²,王鹏³,谢高辉⁴

¹ 中国科学院成都计算机应用研究所，四川成都 610041
² 中国科学院大学，北京 100049
³ 西南民族大学计算机科学与技术学院，四川成都 610041
⁴ 广州五舟科技股份有限公司，广东广州 510000

出版日期:2016-03-25 发布日期:2017-08-04
基金资助:
国家自然科学基金资助项目;广东省科技厅高新技术产业化科技攻关基金资助项目;成都市创新发展战略研究基金资助项目

Optimizing demand charge of data center base on PE method

Yan HUANG^1,²,Peng WANG³,Gao-hui XIE⁴

¹ Chengdu Institute of Computer Application,Chinese Academy of Sciences,Chengdu 610041,China
² University of Chinese Academy of Sciences,Beijing 100049,China
³ School of Computer Science and Technology,Southwest U iversity for Nationalities,Chengdu 610041,China
⁴ Guangzhou Wuzhou Technology Corporation,Guangzhou 510000,China

Online:2016-03-25 Published:2017-08-04
Supported by:
The National Natural Science Foundation of China;Guangdong Science and Technology Plan of Indus-trial High-Technology;Chengdu Innovation and Development Strategy Research Project

摘要/Abstract

摘要：

数据中心的电费包含需量费用和用量费用2个部分，以往的数据中心电费优化研究通常未考虑需量费用。在满足服务质量和响应时间的约束条件下采用PE(partial execution)方法可以减少峰值功率，从而降低需量费用。通过对时间、外部负载、服务质量约束、响应时间约束进行建模，构建电费优化问题的模型，提出基于 PE 方法的需量费用优化算法PEDC（partial execution demand charge)，在高负载时间段执行PE方法减小峰值功率以降低需量费用并降低用量费用，从而降低数据中心的总电费。通过与4种算法进行比较，并构造准确预测的负载数据对PEDC算法性能进行测试，PEDC算法可以降低数据中心运营总电费5.9%~12.7%，提升集群使用率1.32倍。

关键词: PE方法, 需量费用, 用量费用, 能耗费用优化

Abstract:

Demand charge and energy charge are the two main components of data center electricity cost,previous re-searches have not take demand charge into consideration.PEDC algorithm was proposed by modeling time slot,work-load,service quality constraint and response time constraint.With PEDC algorithm peak power was decreased by partial execution on the condition of service quality constrai and response time constraint.PE method was executed in the heavy loaded time slots to reduce peak power so as to ize demand charge.Energy charge and total charge were also optimized.By comparing with four algorithms and with accurately predicted,PEDC algorithm can reduce elec-tricity cost by 5.9%~12.7% and improve cluster utilization 1.32 times.

Key words: PE method, demand charge, energy charge, energy cost optimization

黄焱,王鹏,谢高辉. 基于PE方法的数据中心需量费用优化算法[J]. 通信学报, 2016, 37(3): 90-97.

Yan HUANG,Peng WANG,Gao-hui XIE. Optimizing demand charge of data center base on PE method[J]. Journal on Communications, 2016, 37(3): 90-97.

图/表 8

表1

图1

图2

图3

图4

图5

图6

图7

参考文献 26

[1]	[EB/OL]..
[2]	[EB/OL]..
[3]	[EB/OL]..
[4]	GAO PX , CURTIS AR , WONG B , et al. It's not easy being green[J]. ACM SIGCOMM Computer Communication Review, 2012,42(4): 211-222.
[5]	HAMILTON J . Cooperative expendable micro-slice servers(CEMS):low cost,low power servers for internet-scale services[C]// CIDR'09(DSN). c2009.
[6]	BELOGLAZOV A , BUYYA R , LEE YC , et al. A taxonomy and sur-vey of energy-efficient data centers and cloud somputing systems[J]. Advances in Computers, 2011,82(2): 47-111.
[7]	BARROSO LA , H?LZLE U . The case for energy-proportional com-puting[J]. IEEE Computer, 2007,40(12): 33-37.
[8]	YAO Y , HUANG L , SHARMA A , et al. Data centers power reduction:a two time scale approach for delay tolerant workloads[J]. IEEE IN-FOCOM, c2012:1431-1439.
[9]	GANDHI A , HARCHOL M , DAS R , et al. Optimal power allocation in server farms[J]. ACM SIGMETRICS, c2009:157-168.
[10]	LIN M , WIERMAN A , ANDREW LL , et al. Dynamic right-sizing for power-proportional data centers[J]. IEEE/ACM Transactions on Net-working, 2013,21(5): 1378-1391.
[11]	RAO L , LIU X , XIE L , et al. Minimizing electricity cost:optimization of distributed internet data centers in a multi-electricity-market envi-ronment[J]. IEEE INFOCOM, c2010:1-9.
[12]	RAO L , LIU X , ILIC MD , et al. Distributed coordination of internet data centers under multiregional electricity markets[J]. Proceedings of the IEEE, 2012,100(1): 269-282.
[13]	LI J , LI Z , REN K , et al. Towards optimal electric demand manage-ment for internet data centers[J]. IEEE Transactions on Smart Grid, 2012,3(1): 183-192.
[14]	LUO J , RAO L , LIU X . Temporal load balancing with serv ce delay guarantees for data center energy cost optimization[J]. IEEE Transac-tions on Parallel and Distributed Systems, 2014,25(3): 775-784.
[15]	LUO J , RAO L , LIU X . Minimizing electricity cost:optimization of distributed internet data centers in a multi-electricity-market envi-ronment[J]. IEEE INFOCOM, c2013:340-344.
[16]	WANG H , HUANG J , LIN X , et al. Exploring smart grid and data center interactions for electric power load balancing[J]. Performance Evaluation Review, 2014,41(3): 89-94.
[17]	ZHOU Z , FANGMING L , ZONGPENG L , et al. When smart grid meets geo-distributed cloud:an auction approach to datacenter de-mand response[C]// IEEE INFOCOM, c2015.
[18]	ZHANG Y , WANG Y , WANG X . Capping the electricity cost of cloud-scale data centers with impacts on power markets[C]// HPCD'11. c2011.271-272.
[19]	GUO Y , FANG Y . Electricity cost saving strategy in data centers by using energy storage[J]. IEEE Transactions on Parallel and Distributed Systems, 2013,24(6): 1149-1160.
[20]	王巍，罗军舟，宋爱波 . 基于动态定价策略的数据中心能耗成本优化[J]. 计算机学报， 2013,36(3): 599-612. WANG W , LUO J X , SONG A B . Dynamic pricing based energy cost optimization in data center enviroments[J]. Chinese Journal of Computers, 2013,36(3): 599-612.
[21]	RAHMAN A , LIU X , KONG F . A survey on geographic load balanc-ing based data center power management in the smart gr environ-ment[J]. IEEE Communications Surveys ＆ Tutorials, 2014,16(1): 214-233.
[22]	MALTZ DA . Challenges in cloud scale data centers[C]// ACM SIG-METRICS. c2013.3-4.
[23]	HE Y , ELNIKETY S , LARUS J , et al. Zeta:Scheduling interactive services with partial execution[C]// ACM SOCC. c2012.12-25.
[24]	XU H , LI B . Reducing electricity demand charge for data centers with partial execution[C]// ACM e-Energy. c2014.51-61.
[25]	[EB/OL]..
[26]	金嘉晖，罗军舟，宋爱波 . 基于数据中心负载分析的自适应延迟调度算法[J]. 通信学报， 2011,32(7): 47-56. JIN J H , LUO J Z , SONG A B . Adaptive delay scheduling lgorithm based on data center load analysis[J]. Journal on Communications, 2011,32(7): 47-56.

区域	需量价格($/kW)	用量价格($/kWh)	需/用数值比
OR	3.84	0.034 1	112.61
IA	6.26	0.026 44	236.76
OK	10.39	0.021 6	481.02
NC	11.10	0.055 69	199.32
SC	14.76	0.050 37	293.03
GA	16.55	0.005 56	2 976.62

基于PE方法的数据中心需量费用优化算法

Optimizing demand charge of data center base on PE method

在线阅读

PDF下载

可视化

被引次数

摘要/Abstract

引用本文

使用本文

图/表 8

参考文献 26

相关文章 0

Metrics

推荐阅读 0