雲端運算是近年來一直被關注的技術,越來越多的網路應用服務借重雲端運算提供用戶端更加便利的使用經驗。與雲端運算息息相關的系統虛擬化技術,不僅可以提高伺服器的執行效能,也具備動態資料調配的特性。當伺服器因為設備故障、負載過大或維護等因素需要停止運轉時,可透過即時遷移技術將資料快速的轉移到備援伺服器上,且使用者不會察覺到服務中斷。 本篇論文介紹在即時轉移技術中,透過預測記憶體修改機率,判斷頁面是否進入暫停-複製階段(Stop-and-copy Phase),減少資料在網路上做不必要的轉移,進而縮短資料遷移的時間。 先前研究透過其他方式,進行記憶體修改率預測,但必須統計到第十次疊代頁面的歷史變化狀態,才能夠準確預測記憶體修改機率。本研究使用Game Theory模型,只需知道本次與前一次的記憶體修改機率,可有效降低預測次數,提早判定頁面是否進入Stop-and-copy階段,進而節省即時遷移所需的時間。
Cloud computing was a technology in recent years which had been concerned. More and more network applications provided client a more convenient experience for use on the cloud computing service. Cloud computing is using virtualization technology. It can not only improve the performance on the server, but including a characteristic dynamic data assignment. Additionally, any server with fault, over loading or maintenance…etc. which need to be stopped, the user is not aware that the service has interrupted, that is because the technology of live migration will quickly backup the remaining data from original server to another server. This study introduces that predicting the probability of memory modification to determine the dirty page whether to go the Stop-and-copy phase in the technology of live migration. This method can avoid data be unnecessary transferred over the network, and reduce the time of migration. The previous studies used other methods to predict the probability of memory modification. But it must be statistics the historical changes of ten times iteration, to be able to accurately predict the probability of memory modification. This study used Game Theory model has a capability just to know this and previous probability of memory modification. It can effectively reduce the predicted numbers to determine the dirty page whether to go Stop-and-copy phase, thus saving time required for live migration.