Yijia Zhang , Athanasios Tsiligkaridis , Ioannis Ch. Paschalidis , Ayse K. Coskun
{"title":"Data center and load aggregator coordination towards electricity demand response","authors":"Yijia Zhang , Athanasios Tsiligkaridis , Ioannis Ch. Paschalidis , Ayse K. Coskun","doi":"10.1016/j.suscom.2024.100957","DOIUrl":null,"url":null,"abstract":"<div><p><span><span><span>In a demand response scenario, coordinating multiple data centers<span> with an electricity load aggregator provides opportunities to minimize electricity cost and absorb the volatility in the grid that is caused by </span></span>renewable generation<span>. To enable optimal coordination, this paper introduces a joint data center and aggregator optimization framework that minimizes the cost of data centers while they participate in </span></span>demand response programs regulated by a load aggregator. The proposed framework, </span><em>DCAopt</em><span><span>, solves three integrated optimization problems: optimizing the quality-of-service of jobs in each data center, coordinating workload sharing among multiple data centers, and assigning (electricity) prices that incentivize demand response. Instead of relying on simplified relations between a data center’s overall utilization rate and the average job delay, DCAopt applies </span>queueing theory<span> and job scheduling simulation techniques to model data centers with heterogeneous workloads, where different workload properties can be measured using data from actual servers. DCAopt solves the aforementioned joint optimization problems via gradient descent. Through evaluation using fine-grained simulations, we demonstrate that our framework finds better solutions to the data-center-aggregator optimization problems. With DCAopt, the energy costs of data centers can be reduced by 5% on average, with a corresponding reduction of a social cost assessed by the aggregator amounting to more than 30% in most cases. In addition, power usage reduction at the data centers is 6% higher compared to data-center-centric power use optimization.</span></span></p></div>","PeriodicalId":48686,"journal":{"name":"Sustainable Computing-Informatics & Systems","volume":"42 ","pages":"Article 100957"},"PeriodicalIF":3.8000,"publicationDate":"2024-01-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Sustainable Computing-Informatics & Systems","FirstCategoryId":"94","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S2210537924000027","RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"COMPUTER SCIENCE, HARDWARE & ARCHITECTURE","Score":null,"Total":0}
引用次数: 0
Abstract
In a demand response scenario, coordinating multiple data centers with an electricity load aggregator provides opportunities to minimize electricity cost and absorb the volatility in the grid that is caused by renewable generation. To enable optimal coordination, this paper introduces a joint data center and aggregator optimization framework that minimizes the cost of data centers while they participate in demand response programs regulated by a load aggregator. The proposed framework, DCAopt, solves three integrated optimization problems: optimizing the quality-of-service of jobs in each data center, coordinating workload sharing among multiple data centers, and assigning (electricity) prices that incentivize demand response. Instead of relying on simplified relations between a data center’s overall utilization rate and the average job delay, DCAopt applies queueing theory and job scheduling simulation techniques to model data centers with heterogeneous workloads, where different workload properties can be measured using data from actual servers. DCAopt solves the aforementioned joint optimization problems via gradient descent. Through evaluation using fine-grained simulations, we demonstrate that our framework finds better solutions to the data-center-aggregator optimization problems. With DCAopt, the energy costs of data centers can be reduced by 5% on average, with a corresponding reduction of a social cost assessed by the aggregator amounting to more than 30% in most cases. In addition, power usage reduction at the data centers is 6% higher compared to data-center-centric power use optimization.
期刊介绍:
Sustainable computing is a rapidly expanding research area spanning the fields of computer science and engineering, electrical engineering as well as other engineering disciplines. The aim of Sustainable Computing: Informatics and Systems (SUSCOM) is to publish the myriad research findings related to energy-aware and thermal-aware management of computing resource. Equally important is a spectrum of related research issues such as applications of computing that can have ecological and societal impacts. SUSCOM publishes original and timely research papers and survey articles in current areas of power, energy, temperature, and environment related research areas of current importance to readers. SUSCOM has an editorial board comprising prominent researchers from around the world and selects competitively evaluated peer-reviewed papers.