公开(公告)号：US09563216B1

公开(公告)日：2017-02-07

申请号：US14084835

申请日：2013-11-20

Applicant: Google Inc.

Inventor： Luiz Andre Barroso ,  Christopher G. Malone ,  Taliver Brooks Heath ,  Nathaniel Edward Pettis ,  Stephanie Hua Taylor ,  Michael C. Ryan

IPC: G06F19/00 ,  G05F1/66

CPC classification number: G05F1/66

Abstract: Techniques for managing power loads of a data center include electrically coupling a data center infrastructure power load and a data center IT power load in a power distribution system having a specified power capacity, the infrastructure power load including a plurality of infrastructure power loads associated with at least one of a data center cooling system, a data center lighting system, or a data center building management system, and the IT power load including a plurality of IT power loads associated with a plurality of rack-mounted computing devices; determining that a predicted amount of the IT power load is about equal to or greater than a threshold power value; throttling the infrastructure power load to reduce a portion of the power capacity used by the infrastructure power load; and based on throttling the infrastructure power load, increasing another portion of the power capacity available to the IT power load.

Abstract translation: 用于管理数据中心的电力负载的技术包括将数据中心基础设施电力负载和数据中心IT电力负载电耦合到具有指定功率容量的配电系统中，该基础设施电力负载包括与之相关联的多个基础设施电力负载 数据中心冷却系统，数据中心照明系统或数据中心建筑物管理系统中的至少一个以及包括与多个机架式计算设备相关联的多个IT电力负载的IT电力负载; 确定IT功率负载的预测量大约等于或大于阈值功率值; 节省基础设施电力负载，以减少基础设施电力负荷使用的一部分电力容量; 并且基于节省基础设施电力负载，增加IT电力负载可用的电力容量的另一部分。

公开(公告)号：US20180204116A1

公开(公告)日：2018-07-19

申请号：US15410547

申请日：2017-01-19

Applicant: Google Inc.

Inventor： Richard Andrew Evans ,  Jim Gao ,  Michael C. Ryan ,  Gabriel Dulac-Arnold ,  Jonathan Karl Scholz ,  Todd Andrew Hester

IPC: G06N3/08 ,  G06N3/04

CPC classification number: G06N3/0454

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for improving operational efficiency within a data center by modeling data center performance and predicting power usage efficiency. An example method receives a state input characterizing a current state of a data center. For each data center setting slate, the state input and the data center setting slate are processed through an ensemble of machine learning models. Each machine learning model is configured to receive and process the state input and the data center setting slate to generate an efficiency score that characterizes a predicted resource efficiency of the data center if the data center settings defined by the data center setting slate are adopted t. The method selects, based on the efficiency scores for the data center setting slates, new values for the data center settings.