Managing the complex data center environment: an Integrated Energy-aware Framework
Managing the complex data center environment: an Integrated Energy-aware Framework
复制标题
DOI：
10.1007/s00607-014-0405-x
复制
发表时间：
2014-05
期刊：
影响因子：
3.7
通讯作者：
Alexandre Mello Ferreira;B. Pernici
中科院分区：
文献类型：
--
作者：
Alexandre Mello Ferreira;B. Pernici
由于企业和个人对信息技术的使用不断增加，信息技术的能源消耗问题已经引起了人们的广泛关注。特别是，数据中心现在在现代社会中发挥着越来越重要的作用，信息随时随地都可以获得。在此背景下，本文的目的是从信息系统的角度研究数据中心内的能效问题。拟议的方法集成了应用程序和基础设施能力，其中适应机制的制定与业务流程保持一致。基于服务应用的能量和质量维度，基于模型的方法支持建立新的约束优化问题，该问题考虑了过约束解，其中的目标是在能量和质量要求之间获得更好的权衡。这些想法结合在一个框架内，在这个框架中，基于时间的分析可以确定潜在的系统威胁，并推动选择适应行动，以改善以指标满意度为代表的总体能源和质量要求。此外，该框架还包括一个演化机制，该机制能够评估过去的决策反馈，以便根据当前的底层环境调整模型。最后，在实验环境中分析了该方法的优点。
The problem of Information Technology energy consumption has gained much attention due to the always increasing use of IT both for business and for personal reasons. In particular, data centers are now playing a much more important role in the modern society, where the information is available all the time and everywhere. In this context, the aim of this paper is to study energy efficiency issues within data centers from the Information System perspective. The proposed approach integrates the application and infrastructure capabilities, in which the enactment of adaptation mechanisms is aligned with the business process. Based on both energy and quality dimensions of service-based applications, a model-based approach supports the formulation of new constrained optimization problem that takes into consideration over-constrained solutions where the goal is to obtain the better trade-off between energy and quality requirements. These ideas are combined within a framework where time-based analysis allow the identification of potential system threats and drive the selection of adaptation actions improving overall energy and quality requirements, represented by indicators satisfaction. In addition, the framework includes an evolution mechanism that is able to evaluate past decisions feedback in order to adjust the model according to the current underlying environment. Finally, the benefits of the approach are analyzed in an experimental setting.