CSR: Small: Understanding and Modeling the Trade-Offs in Data Centers for Next-Generation Sustainable Management

CSR：小：理解和建模数据中心的权衡以实现下一代可持续管理

• 批准号: 1218505
• 负责人: Sandeep Gupta
• 金额: $ 45万
• 依托单位: Arizona State University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2012
• 资助国家: 美国
• 起止时间: 2012-08-01 至 2017-07-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1218505&HistoricalAwards=false
• 关键词: CSR; Small; Understanding; Modeling; Trade

The success of cloud computing as a business model has led to several commercial cloud services for storage, data mining and searching, and multimedia content delivery. As cloud services penetrate emerging markets -- notably the smart-phone market -- conventional utility services such as television are becoming ?cloudized.? As the economy becomes more dependent on the efficient operation of cloud data centers (DCs), there?s an increasingly urgent need to design ?green? and energy-adaptive management software that optimizes the operation of a geographically distributed set of data centers under varying workload and energy conditions.Cyber-physical, proactive and predictive decision making is essential for a DC operating system that efficiently predicts, e.g., the workload and the power consumption when running multiple jobs on the same server and makes decisions, e.g., for redistributing the servicing virtual machines among varying number of available DCs, electricity cost, percentage of renewable energy supply and capacity of energy storage. This project is addressing the scientific and engineering challenges involved in decision making of such DC management arising from an array of complicating factors -- e.g., server heterogeneity, cyber-physical interactions between DCs and their environment, varying workload and electricity pricing -- and trade-offs -- e.g., energy-delay, time-accuracy, and cooling-computing . The project is developing a Green cyber-physical Data Center Simulator, GDCSim, for studying the tradeoffs in practical scenarios with several complicating factors. Project aims to enable a more sustainable cloud computing and to train engineers for a greener economy. The project findings are available through ASU-IMPACT lab?s website (http://impact.asu.edu/).

云计算作为一种商业模式的成功催生了多种用于存储、数据挖掘和搜索以及多媒体内容交付的商业云服务。随着云服务渗透到新兴市场——尤其是智能手机市场——电视等传统公用事业服务正在变得“云化”。 随着经济越来越依赖云数据中心 (DC) 的高效运行，“绿色”设计的需求也越来越迫切。能源自适应管理软件，可在不同的工作负载和能源条件下优化地理上分布的一组数据中心的运行。网络物理、主动和预测性决策对于有效预测工作负载和能源等问题的数据中心操作系统至关重要。在同一服务器上运行多个作业并做出决策时的功耗，例如，在不同数量的可用 DC 之间重新分配服务虚拟机、电力成本、可再生能源供应百分比和能源存储容量。该项目正在解决此类数据中心管理决策中涉及的科学和工程挑战，这些挑战是由一系列复杂因素引起的，例如服务器异构性、数据中心及其环境之间的网络物理交互、不同的工作负载和电价以及贸易-offs——例如能量延迟、时间精度和冷却计算。该项目正在开发绿色网络物理数据中心模拟器 GDCSim，用于研究具有多个复杂因素的实际场景中的权衡。 该项目旨在实现更可持续的云计算，并培训工程师以实现更绿色的经济。 该项目的研究结果可通过 ASU-IMPACT 实验室网站 (http://impact.asu.edu/) 获取。