隨著數位科技快速發展，資訊服務的持續可用性已成為企業營運成功的關鍵因素之一。高可用性（High Availability, HA）架構可確保資訊系統在面臨硬體故障、軟體錯誤或災難事件時，仍能穩定運作，進而降低業務中斷風險與系統停擺所帶來的經濟損失。隨企業全球化布局日益擴大，對跨地域資訊服務不中斷的需求愈加迫切，如何縮短系統異常後的恢復時間，亦成為資訊化策略的重要挑戰。本研究目的在提升企業資訊系統之可用性與營運持續性，設計並建構一套支援跨廠區部署之超融合基礎架構（Hyper-Converged Infrastructure, HCI），整合 VMware vSphere 虛擬化平台與 StorMagic SvSAN 軟體定義儲存（Software-Defined Storage, SDS）技術。透過導入第三方仲裁節點與跨廠區資料同步機制，有效預防雙節點架構中常見之裂腦（Split-Brain）問題，並確保系統資料之一致性與即時性。研究中於實際企業場域建置此高可用性架構，並透過模擬主機故障、網路中斷及裂腦發生等情境，驗證系統於異常發生時能在 99 至 123 秒內自動完成容錯切換與服務恢復。實驗結果顯示，所建構之架構具備完善的故障偵測、自動接管與儲存一致性控制機制，能有效降低營運中斷風險，並提升企業資訊系統之韌性與穩定性。論文除詳述說明其架構設計理念與實作流程，亦透過多項異常情境測試，針對其適用條件進行實證分析，驗證其於實務環境中之可行性與應用效益。;With the rapid advancement of digital technology, the continuous availability of information services has become a critical factor for enterprise success. High Availability (HA) architectures ensure systems remain operational during hardware failures, software errors, or disasters, reducing risks of downtime and business loss. As global operations expand, enterprises face increasing demand for uninterrupted cross-site services and rapid recovery from failures. This study aims to enhance system availability and continuity by designing a cross-site Hyper-Converged Infrastructure (HCI) integrating VMware vSphere and StorMagic SvSAN Software-Defined Storage (SDS). A third-party witness node and real-time data synchronization are introduced to prevent split-brain issues and ensure data consistency. The architecture was deployed in a real enterprise environment and validated through simulated host failures, network disconnections, and split-brain scenarios. Results showed that automated failover and service recovery completed within 99 to 123 seconds. The proposed solution provides robust fault detection, automated takeover, and consistent storage control, significantly improving system resilience and stability. This research confirms the architecture’s practical feasibility and its value in supporting continuous enterprise operations.
