隨著雲端運算與虛擬化技術的普及，利用虛擬機器提供服務，將多個不同伺服器或服務整合至一台實體機器上，以彈性分配和利用硬體資源，但相對於實體機器，虛擬機更易因硬體設備或人為因素發生故障，造成服務中斷，因此必須使用高可用性技術，確保其可用性，基於虛擬化技術的容錯系統能在虛擬機器因某些因素發生故障時保持虛擬機之運行狀態，使服務不中斷，提升系統可用性。本研究基於中央大學平行與分散計算實驗室開發的NCU MFTVM容錯系統上繼續進行開發。NCU MFTVM屬於持續同步架構，持續同步技術不間斷地將主要虛擬機的狀態同步到備援虛擬機上，可透過外部管理介面控制的容錯系統，並基於容錯系統的準則，保持兩者對外輸出的一致性。本研究在虛擬機容錯系統中加入Intel® QAT硬體壓縮功能，改善了容錯系統的壓縮機制，在虛擬機容錯系統在同步大量資料時，降低NCU MFTVM容錯系統的回應時間，提升容錯系統的可靠性以及可用性。;As cloud computing and virtualization technologies become more widespread, utilizing virtual machines to provide services and consolidating multiple different servers or services onto a single physical machine allows for flexible allocation and utilization of hardware resources. However, compared to physical machines, virtual machines are more prone to failures due to hardware issues or human errors, leading to service corruption. Therefore, it is necessary to employ high availability technologies to ensure their availability. Fault-tolerant systems based on virtualization technology can maintain the operational state of virtual machines when they encounter failures, preventing service disruptions and enhancing system availability. Our research is based on the NCU MFTVM fault-tolerant system, which belongs to a continuous checkpointing architecture and can be controlled through an external management interface. Continuous checkpointing ensures constant synchronization of the state from the primary virtual machine to the backup virtual machine, adhering to fault-tolerant system criteria to maintain output consistency across both VMs. Ours research incorporates Intel® QAT hardware compression feature into the virtual machine fault-tolerant system, improving the compression mechanism of NCU MFTVM fault-tolerant system and reducing the response time during the synchronization of large amounts of data, thereby enhancing both its reliability and availability.
