近年來隨著網路應用服務越來越多元化，自從雲端服務開始廣為被商業市場接受與運用後，許多企業逐漸開始重視內部資料中心基礎設施的佈建以及運算需求，不再使用過去有需要才購買相關硬體設備的方式，而是在建置前就先開始妥善規劃運算資源與應用服務需求，並開始導入虛擬化技術徹底發揮伺服器運算能力，以利達到運算彈性拓展、服務快速部署、資源有效利用三大需求。現今企業為避免過去資料中心硬體規格、作業系統與應用服務都不盡相同的狀況，在資料中心設計上已盡可能使用相同的硬體設備與作業系統來提供服務，以降低維運管理中的各種不確定性與複雜程度。資料中心為滿足雲端服務的應用，在部署作業必須採取更彈性且快速的設計與規劃思維。過去在資料中心的部署作業，往往需要人力適時介入操作系統與設定環境參數，方可完成伺服器的部署程序；倘若遇上伺服器故障問題，必須要花費大量時間才可能修復再重新上線，以傳統部署作業方式，易造成無法即時回應運算需求的瓶頸，也造成資料中心人員效率不彰與即時反應能力不佳。本研究發現市場上雖然有自動化部署系統的解決方案，但無法達到真正自動化作業，仍然需要專業且熟悉的資訊人員才能得以操作與設定，雖然這類部署系統比起作業系統所提供的自動安裝程式(Kick-Start)方便許多，但在資料中心的維運管理上，企業勢必要投入相對應的資源才能支撐自動化部署系統的運作。本研究針對近年相當熱門的軟體定義特性用於資料中心的自動化部署系統相關議題進行整理與分析，同時實作一個基於軟體定義的自動化部署系統來驗證此系統可行，論文中也對系統相關的運作效能進行分析與說明。;Since cloud services become the trend and accepted by the commercial market, more and more network application services comes up. Now, many enterprises have gradually begun to pay attention provisioning and operational needs of the internal data center infrastructure. They like to plan and forecast needs and resource rather than purchase hardware when needs comes up. Enterprises also use virtualization technology in different parts like maximize server’s computing power, fast deploy of service and efficient of resource. It’s a big problem and challenge if hardware, OS and application is all different in each system of enterprise. Data center has been designed as far as possible using the same hardware and operating system to provide services to reduce the degree of uncertainty and complexity in a variety of maintenance and operation management. Enterprise wish Data center can fulfill cloud service’s needs, the deployment must adopt more flexible, rapid design and forecast plan. In the past, deployment of date center, it often require timely intervention of human operating system and set of environmental parameters to complete server’s deployment process. If server have some problem, it will take a lot of time to recover if we didn’t have automation deploy. It will also cause scale up problem, maintain and responds time slowly problem. Although the market has automation deployment system solutions, but cannot achieve true automation. It still require professional and hands on operator to do configuration and operation. Those kind of automation deployment system solutions is more convenient than OS’s Kick-Start. Even if so, it’s still require lots of resource to let that kind of automation works. It’s the hot topic of information technology which call “Software-Defined”. This study is about the how to use software-defined into data center automation deployment system. To implement a software-defined based automation deployment system to verify the feasibility of this system. This study is also related to the operation of the system performance analysis and explanation
