實作一個資料中心的雲端綠能管理系統

本研究提出建置一個資料中心的綠能管理系統包含用電分析與處理服務，於本校電子計算機中心主機房建置系統，以降低整體用電量與PUE為目標。我們以數位電錶對用電設備進行資料收集，並搭配環境溫濕度與CO2資訊，提供使用者以圖型化界面呈現以利資料分析。透過系統紀錄與用電分析，可隨時隨地以行動裝置對用電設備進行監看。運用巨量資料處理與分析架構為基礎，建置一個用電端綠能管理需求為核心的應用服務平台。建置Hadoop架構來實作整體運算及電能資料的儲存，利用MapReduce的運算架構，透過分散式的運算方式來存取HBase，來達到巨量資料快速的資料分析及處理，用以提高資料處理效能。提供龐大的資料存取與分析，由電能監控取得的相關資訊有效估算用電量，讀取用電的歷史紀錄來分析用電行為，找出可以改善的地方，以達有效降低電能的浪費。

關鍵字

巨量資料；雲端運算； Hadoop ；資料分析；電能監控

並列摘要

In our research, we provide a Green Energy Management System of data center in the Computer Center of Tunghai University that including power usage analysis and processing services. The main goal of the system is to reduce the whole consumption and PUE. In order to supply a graphical user interface that can easily analyze data, we use digital meters to collect data from equipment and collocation with the information of temperature, humidity, and CO2. By analyzing system record and power usage, we can monitor our equipment by mobile equipment in anytime and anywhere. Based on Big Data Processing and analysis architecture, implementing an application service platform as a kernel of the demands of cloud green management. Providing a great amount of data access and efficiently analysis estimate power usage for energy monitoring information. After reading historical record to analyze the habit of power usage, we can figure out the weakness of data center environment that can be improved, to reach the goal of reducing power waste. We use Hadoop computing architecture to implement our system including computing and storage. In the energy data storage part, achieving rapid analysis and processing of big data by use Hadoop MapReduce architecture access HBase.

並列關鍵字

Big Data ； Cloud Computing ； Hadoop ； Data Analysis ； Energy Monitoring ； Hadoop MapReduce

國際替代計量

實作一個資料中心的雲端綠能管理系統

全文下載

主題瀏覽