標題: | 利用SDN技術增進OpenStack網路使用率 Using SDN Technology to Improve the Network Utilization of OpenStack |
作者: | 張華婷 Chang, Hua-Ting 王協源 王豐堅 Wang, Shie-Yuan Wang, Feng-Jian 網路工程研究所 |
關鍵字: | OpenStack;SDN;Live-migration;OpenStack;SDN;Live-migration |
公開日期: | 2013 |
摘要: | 由於網路的快速發展以及雲端服務的普及化,資料中心的內部架構越來越受到重視,而在資料中心內部的網路壅塞問題更是不能小覷,若不能有效的減緩網路壅塞問題,將會嚴重影響到在資料中心上所運行的應用和服務。
本論文將使用SDN技術和OpenStack所提供的虛擬化技術和控管功能,提出一套機制來控制OpenStack內部的網路,在發生網路壅塞時,本系統會利用Virtual Machine (VM)的live-migration和流量路徑調整來紓解網路中的壅塞,並在VM live-migration後配合SDN的技術避免傳統網路中的拓樸收斂時間,以達到提升網路使用率。
為了驗證本論文所提出的方法,我們針對本系統的效能和目前廣泛在資料中心使用的Equal-cost multi-path routing (ECMP)的效能做了不同流量型態下的實驗,評比結果顯示不論是固定速度的流量或不限速的流量都能提升網路的使用率;除此之外,也在實驗中發現VM live-migration配合SDN的做法比改變路徑更能有效的紓解網路壅塞的情形。 Owing to rapid development of the Internet and the popularization of cloud services, the internal structure of the data center has received much more attention. As for the network in the data center, the importance of solving network congestion cannot be overemphasized. The network congestion, which can’t be solved efficiently, may influence users, who use the cloud services. A solution, which is related to SDN technology, solving the network congestion in the data center is introduced in this thesis. When network congestion takes place, the VM (virtual machine) live-migration design, which cooperates with SDN to avoid the extra convergence time triggered by topology changing, and the path adjustment path are used to relieve the network congestion and improve the network utilization of OpenStack. We performed several experiments under the fixed-rate flows and the unlimited-speed flows conditions, and then compared the results of the proposed method to the results of ECMP. Finally, we showed that the proposed method can improve the network utilization of OpenStack. Besides, we also discovered that the VM live-migration design contributes more than the design of changing paths of flows. |
URI: | http://140.113.39.130/cdrfb3/record/nctu/#GT070156524 http://hdl.handle.net/11536/75047 |
Appears in Collections: | Thesis |