云计算资源占用量化评估体系的分析与应用-analysis and application of quantitative evaluation system for cloud computing resource occupancy.docxVIP

云计算资源占用量化评估体系的研究与应用摘 要随着互联网应用技术的不断发展，尤其是近年来电子商务、视频、 微博等应用引起的新一轮互联网投资热潮，国内 IDC 业务始终稳步 上升。但与此同时，资源利用率低、能源利用率低、服务层次较低、 传输数据剧增等问题已越来越明显的影响着 IDC 业务进一步发展。 将云计算技术引入 IDC，作为 IDC 增值服务的主要内容之一，将有 效推动运营商解决上述 IDC 业务发展面临的种种问题。然而，相对 于传统 IDC 业务，云计算应用部署较为复杂，需要根据应用系统的 特性，选择合适的部署模式，资源分配和调度模式。为此，本文构建 了云计算资源占用量化评估体系，通过该体系来帮助我们确定应用系 统类型。本文首先对云计算环境下的几种应用部署架构进行了分析，从而 总结出适用于云计算部署的应用特点。其次，对虚拟化环境下，应用 资源占用量化参数和指标进行了分析，总结出 5 种应用资源占用模 式。随后，重点研究利用层次分析法（AHP 方法）构建应用资源占 用量化评估体系。最后，选取 1 个实例对评估体系做了实际应用，验 证了应用资源占用量化体系是合理、有效的。同时，通过对该实例的 测评，证实了利用 AHP 方法所构建的应用资源占用量化评估体系对 电信运营商具有较高的参考价值和应用前景。关键词： IDC，云计算，架构，层次分析法Research and Application of Quantitative Cloud Computing Evaluation System Based Resource ConsumptionABSTRACTWith the development of internet technology, a new round of internet investment starts, including e-business, video websites and micro-blogging etc. This guarantees IDC business a steady grow in China. However, problems like lower in material re-usability, lower in energy utilization, lower in level services and significant increasing amount of transferring data, which are preventing the IDC business from growing. In order to help the operating companies to effectively solve the above problems, we introduce cloud computing technology to IDC and make it part of our major add value business. However, comparing to the traditional IDC business, the deployment of Cloud Computing Application is more complex. We need to choose the proper way to deploy, distribute and distich job based on the characteristic of the system. To make a guideline for identifying the proper system type, we constructed a quantitative Cloud Computing Evaluation System.Firstly, the different types of deployment architecture and summarized the characteristic of Cloud Computing deployment are analyzed in this paper. Secondly, we conclude five resource utilization modals based on the analysis of resource utilization parameters and KPIs from certain application. Thirdly, we used AHP (analytic hierarchy process) to construct quan