成果報告書詳細
管理番号20120000000754
タイトル*平成20年度中間年報 グリーンネットワーク・システム技術研究開発プロジェクト(グリーンITプロジェクト) エネルギー利用最適化データセンタ基盤技術の研究開発 データセンタのモデル設計と総合評価
公開日2012/6/29
報告書年度2008 - 2008
委託先名独立行政法人産業技術総合研究所 国立大学法人筑波大学 エヌ・ティ・ティ・コミュニケーションズ株式会社 ソフトバンクIDC株式会社
プロジェクト番号P08017
部署名省エネルギー部
和文要約和文要約等以下本編抜粋:
1.研究開発の内容及び成果等
(1)サーバシステムの省エネルギー性を評価できる指標の枠組み開発
1【評価指標】サーバシステムの評価指標[担当:筑波大学]
サーバシステムの省エネルギー性は、単純に定常的な消費電力量だけでは測ることができず、電力を供給することによって可能となるサーバの性能やサービスの質を加味する必要がある。性能を加味した指標を開発するため、サーバシステムのコンポーネント毎の消費電力が把握可能か調査を行うとともに、省エネルギー性を評価できる指標の枠組みを検討している。当該年度においては、サーバシステムのコンポーネント毎の消費電力が把握可能か調査した。
英文要約Title: Research and Development Project for Green Network/System Technology (Green IT Project). R&D of Fundamental Technology for Data Centers to Optimize Energy Use/ Model Designs and a Comprehensive Evaluation of Data Centers (FY2008-FY2009) FY2008 Annual Report
As a preparation of evaluation experiment, we have achieved the following.
(1) Development of new metrics for evaluating energy efficiency of server system (University of Tsukuba)
We are investigating new metrics for evaluating the energy efficiency of server systems, firstly by exploiting the way to measure the power consumption of each component and configuration of the server system. We have set an experimental target system and the measurement system up, and started the measurement. We need an additional power measurement system to measure the power of each component, but some recent systems are already equipped with the power monitoring system. We considered metrics to evaluate the energy efficiency of server systems, and found it difficult to take factors such as the capacity and QoS into account. We need some metrics to reflect the performance and QoS of server systems. We will make the relationship between these factors and powers clear for design a new metrics of the energy efficiency.
(2) Development of new metrics for evaluating energy efficiency in a data center (AIST)
We have developed framework of new metrics for evaluation of energy efficiency in a data center. The metrics include the productivity of the data center and the power consumption. However we need more investigation of business application in data centers in order to define the productivity clearly. Furthermore, to construct the experiment environment we introduced IT devices, measurement devices and developed the software which are used to extract required resources. We will complete system for the experiment environment and use it for modeling data center and developing new metrics.
(3) Investigation of existing metrics for evaluating energy efficiency (NTT communications)
We have researched existing metrics for evaluating energy efficiency and listed their problems and subjects up. There are several existing metrics, such as PUE, which are well used. However, most of them are not affected by newly developed "energy saving technologies". Therefore, metrics for evaluating energy efficiency should have new parameters. Furthermore, there are environmental factors which affect the energy consumption. The environmental factors should be defined for the measuring parameters of metrics.
(4) Preliminary investigation for evaluation experiments (Softbank IDC)
We researched measurement ways of the power consumption and possible restrictions of data center for evaluation experiments. We obtained three pieces of knowledge. First, there are several types of data centers which can be distinguished by facilities and business operations. Second, there are two ways to measure the power consumption of the IT devices in rack. Last, there are restrictions to set facilities in data center. For example, the facilities which will be developed by other project teams may be too big or too heavy. The old type of data center may not accept them.
ダウンロード成果報告書データベース(ユーザ登録必須)から、ダウンロードしてください。

▲トップに戻る