| 摘要: | 近年雲端服務的應用逐漸普及到企業內部,越來越多的企業開始將公司內部的服務轉移到
私有雲或是公有雲上面,而混和使用公有雲及私有雲是常見的雲端基礎建設需求,通常在私有
雲上面的服務,大多是重要且不能中斷的服務,只要發生重大故障導致服務不能運行,企業都
會蒙受重大的損失,故雲端平台的可用性 (Availability) 將會變得非常重要,不論是針對系統
設備的意外狀況處理,或是維護時的服務轉移,可用性相關的技術都能夠提供許多的幫助,讓
企業在面對大型基礎建設的災害時不會顯得措手不及。
本研究主要基於國立中央大學平行與分散計算實驗室所開發的 NCU MFTVM 容錯系統進
行後續的開發,該容錯系統亦屬於面相雲端服務高可用性的一種技術,而本研究則針對原先的
功能進行優化與改善,針對虛擬硬碟同步技術的代價進行優化,以及改善長久開發所累積的技
術債,針對系統架構進行部分改良,藉此降低後續開發人員進行開發的負擔。;In recent years, more and more enterprises migrate their services to Cloud platform. And
a hybrid Cloud platform is a popular option for enterprises to build their own infrastructure.
Basically, their will deploy their core services on private cloud but not public cloud. When
any services on a cloud are unavailable, the enterprise may lose a bundle before all services are
recovery. Therefore, the availability of cloud platform become more important when services
is used on business.
This research is based on the NCU MFTVM fault-tolerant system developed by the Paral-
lel and Distributed Computing Laboratory of National Central University. We develop a new
method to synchronize block with a lightweight mechanism which can reduce the time we need
to create a new checkpoint. And we also reengineer a part of our fault-tolerant system to improve
the readability and the maintainability. |
