博碩士論文 100522002 完整後設資料紀錄

DC 欄位 語言
DC.contributor資訊工程學系zh_TW
DC.creator洪源懋zh_TW
DC.creatorYuan-Mao Hongen_US
dc.date.accessioned2013-8-28T07:39:07Z
dc.date.available2013-8-28T07:39:07Z
dc.date.issued2013
dc.identifier.urihttp://ir.lib.ncu.edu.tw:88/thesis/view_etd.asp?URN=100522002
dc.contributor.department資訊工程學系zh_TW
DC.description國立中央大學zh_TW
DC.descriptionNational Central Universityen_US
dc.description.abstract近幾年雲端運算(Cloud Computing)的興起,對ISP業者而言,來自使用者的流量與虛擬運算節點擴增的需求都大量的提升。但在現今的網路架構中,各設備皆獨具一格,其上有各自的作業系統與依附的應用程式,故難以因應現有網路架構下,在其上開發符合管理者需求的雲端應用程式。OpenFlow是作為實現具體實踐軟體定義網路(Software Defined Networking;SDN)願景的一開放式網路架構,目的是為了解決傳統交換器與路由器等網路設備因其作業系統與應用程式各自獨立,網路管理者無法因應自身需求來調整、管理網路資料的傳遞路徑等問題。故希望透過OpenFlow可集中式網管、快速佈署符合本身需求的應用程式等能力來增加網路封包路徑的彈性且可進一步將網路管理化繁為簡,提升網路的控制能力、降低網路的複雜度,並自行定義所需的網路。本論文提出智慧型的負載平衡演算法來克服不同flow entry其上負載不平衡的狀況來達到網路的負載平衡(Load balancing),並提供OpenFlow switch與伺服器的備援機制,透過OpenFlow來增加網路資料傳輸的彈性並達到高可用性的系統。 本論文中透過pre-configured flow與dynamic flow來使封包重導至指定的load balancing port上,並透過設置optimal netmask來降低流量項(Flow entry)的數量。另外,在OpenFlow controller上開發負載平衡演算法來決策分流的導向與重導資訊以達負載平衡效果,並利用health check機制來達到高可用性系統。最後,本論文在實驗結果中顯示所提之Genetic-based負載平衡演算法在一般情境會比Load-based及Hash-based負載平衡演算法針對系統負載程度平均後好上9.01%,更比Round-robin負載平衡演算法增進28.25%,顯示本機制更能夠在flow負載不均衡的情形下透過機制計算後進行重導,增進系統的負載平衡程度。zh_TW
dc.description.abstractAs growth of the cloud computing, the traffic flow from clients and the demand for extending virtual nodes are also increasing. But it’s difficult to develop the network applications which match the requirement for network developer because the the applications and operating system for network devices are independent. OpenFlow is a solution for implementing Software Defined Networking(SDN) with the objective to improve the resilience of data transmission. Therefore, we use OpenFlow to achieve network control centrally, deploy applications fast and decrease the complexity for network topology. For these reasons, this thesis used OpenFlow to propose intelligent load balancing algorithms based on getetic algorithm to solve the situation for unbalanced load for each flow entries to achieve traffic load balancing and provide redundant mechanisms to build high availability system. This thesis proposed pre-configured flow and dynamic flow load balancing mechanisms to redirect packet into specified load balancing port to achieve load balancing and set optional netmask to each pre-configured flows in order to reduce the flow rules. Furthermore, this thesis also developed four load balancing algorithms and two health check mechanisms on the OpenFlow controller. At last, the experiments show that genetic-based load balancing algorithm performs 9.01% degree of load balancing better than load-based and hash-based load balancing algorithm, and 28.25% better than round-robin load balancing algorithm. It shows that mechanisms proposed by this thesis can improve the degree of load balancing even the flow entries are unbalanced.en_US
DC.subject雲端運算zh_TW
DC.subject軟體定義網路zh_TW
DC.subjectOpenFlowzh_TW
DC.subject負載平衡zh_TW
DC.subject健康狀態檢測zh_TW
DC.subjectCloud computingen_US
DC.subjectSoftware defined networkingen_US
DC.subjectOpenFlowen_US
DC.subjectLoad balancingen_US
DC.subjectHealth checken_US
DC.title基於OpenFlow的網路負載平衡及高可用性系統zh_TW
dc.language.isozh-TWzh-TW
DC.titleOpenFlow-based Network Load Balancing and High Availability Systemen_US
DC.type博碩士論文zh_TW
DC.typethesisen_US
DC.publisherNational Central Universityen_US

若有論文相關問題,請聯絡國立中央大學圖書館推廣服務組 TEL:(03)422-7151轉57407,或E-mail聯絡  - 隱私權政策聲明