近年來，雲技術已相當成熟，然後透過虛擬化技術可以提高硬件設備的利用度，如何有效地分配和虛擬機的管理，已成為一個非常重要的問題。
在雲計算技術下，多種數據的資料中心為了降低成本並增加系統執行效率使用了的虛擬化技術。然而硬體資源是有限的，隨著資料的增加，系統勢必需要做架構上的更正，畢竟系統為了要長期運作，則系統遷移是必然會發生的問題。
然而，在系統遷移的這段時間，如何在維持正常的系統運作，並完成系統架構遷移是我們實驗說探討問題。
我們使用hypervisor來控制和管理VMs，在系統下為了讓遷移中的系統維持運作，這裡使用了Live Migration的技術。儘管如此，在遷移完成後，兩台VMs的切換依然會讓系統的通訊被迫中斷，然而VMs遷移順序的不同，我們希望讓通訊的品質維持，在此我們透過一些計算，改變了原來排班的結果，讓整個系統通訊能更有效的運作。
我們的系統遷移會在兩個不同的實體機下模擬，在遷移前我們要確保目的端擁有最低限度的資源，讓系統可以正常運作，監控系統VMs遷移的平均利用率時間最高，當然我們希望系統在最短的時間內恢復到準備狀態。也就是說在系統切換的階段，包含了網路設定，讀取服務狀態並繼續執行，在此任務完成前，會讓通訊中斷，所以儘管系統已經完成遷移，但VMs的數據通訊依然閒置。
我們提出了排隊遷移的概念，就是希望在不同的網路拓撲結構下，提供最好的排班方式，為了更清楚的說明演算法的概念，在這裡我們分別測試了四個不同的網路拓撲結構，在後面的章節我們會詳細的說明，最後從我們從設計的系統所獲得的改善結果。

In recent years, cloud technologies have developed well. As a result, we can increase the utilization of hardware virtualization technology. However, how efficiently one can allocate and manage virtual machines (VMs), has become a very important issue.
Under the cloud computing technology, there are a variety of data center in order to reduce cost and to increase the efficiency of the system virtualization technology.
But, as the hardware resources are limited and with the increase of data, the system will need the correct architecture, i.e. especially, in the long-term operation of
the system, there is need of system migration. Thus in this work, we discuss the experimentation of the way to maintain the normal operation of the system by migration, and the complete system architecture for migration.
We use the hypervisor to control and manage VMs of the system, to allow the migration of systems and to maintain the operation we use Live Migration technology. However, after the migration is complete, the two VMs still need to switch so it interrupts the communication. With different VMs migration order,
we still want to maintain the quality of communication. Thus, with the help of few calculations we changed the original scheduling, as a result, the whole system
communication could become more effective in its operation.
Our proposed system will migrate in two different physical machine simulate.
Thus we want to make sure before the migration that the destination has minimal resources, so that the system can operate normally. The average utilization time,
of monitoring system of the VMs migrate to maximum, certainly as we want the system in the shortest possible time to recover to the ready state. That is the stage of system switching, including network settings, service status and to continue reading until this task gets completed, make communication interruption. So even though the system has completed the migration, but still the VMs data communication remains idle.
We proposed the concept of queuing migration, for different network topologies to provide the best scheduling approach. In our proposed algorithm, we tested four different network topologies, and in later chapters we explain them in detail. Finally, we present the results that we obtained from our improved system design.

[1] H. C. Huang and S. Y. Chang. Framework for network virtualization over layer 2. 2014.
[2] S. Jain, A. Kumar, S. Mandal, J. Ong, L. Poutievski, A. Singh, S. Venkata, J. Wanderer, J. Zhou, M. Zhu, et al. B4: Experience with a globally-deployed software defined wan. In ACM SIGCOMM Computer Communication Review, volume 43, pages 3–14. ACM, 2013.
[3] B. Pfaff and B. Davie. The open vswitch database management protocol. 2013.
[4] A. Hayward and S. P. Mullen. Migration of virtual machines, February 28 2013. US Patent App. 13/780,061.
[5] C. K. Ching and M. Martin. An integrated virtual and physical network infrastructure for a networking laboratory. In Computer Science & Education (ICCSE), 2012 7th International Conference on, pages 1433–1436. IEEE, 2012.
[6] A. Kivity, Y. Kamay, D. Laor, U. Lublin, and A. Liguori. kvm: the linux virtual machine monitor. In Proceedings of the Linux Symposium, volume 1, pages 225–230, 2007.
[7] C. Clark, K. Fraser, S. Hand, J. G. Hansen, E. Jul, Christian Limpach, Ian Pratt, and Andrew Warfield. Live migration of virtual machines. In Proceedings of the 2nd conference on Symposium on Networked Systems Design & Implementation-Volume 2, pages 273–286. USENIX Association, 2005.
[8] A. T. Campbell, J. Vicente, and D. A. Villela. Virtuosity: Performing virtual network resource management. In Quality of Service, 1999. IWQoS’99. 1999 Seventh International Workshop on, pages 65–76. IEEE, 1999.
[9] F. Machida, M. Kawato, and Y. Maeno. Just-in-time server provisioning using virtual machine standby and request prediction. In Autonomic Computing, 2008. ICAC’08. International Conference on, pages 163–171. IEEE, 2008.
[10] B. S. Blanchard, W. J. Fabrycky, and W. J. Fabrycky. Systems engineering and analysis, volume 4. Prentice Hall Englewood Cliffs, New Jersey, 1990.