利用平行名稱查詢與貝式分類法增進命名中心網路路由效能

機構典藏 > College of Engineering > Graduate Institute & Department of Electrical Engineering > Thesis > Item 987654321/94526

Please use this identifier to cite or link to this item: https://tkuir.lib.tku.edu.tw/dspace/handle/987654321/94526

Title:	利用平行名稱查詢與貝式分類法增進命名中心網路路由效能
Other Titles:	Enhancing routing efficiency for NDN by parallel name lookup and bayesian classification
Authors:	許惟倫;Hsu, Wei-Lun
Contributors:	淡江大學電機工程學系碩士班吳庭育;Wu, Tin-Yu
Keywords:	命名中心網路;平行名稱查詢;貝式分類法;NDN;CCN;PNL;Bayesian classification
Date:	2013
Issue Date:	2014-01-23 14:44:24 (UTC+8)
Abstract:	命名中心網路(Named Data Networking, NDN)主要是使用以資料名稱命名而不是以現有網路的IP位址來進行資料的傳輸、建立網路路由資料暫存庫與基於內容命名的路由機制，來降低資料傳輸的流量，加快資料獲取的相對應速度。在命名中心網路當中，網路可以辨別資料內容，並且將他們儲存到最近的路由器上，藉此提供從用戶端到最近的資料內容的最佳化路徑，透過最佳化的路徑和資源來確保資料內容傳輸的高效性。命名中心網路(NDN)是下一代網路中熱門的議題之一，而針對於命名中心網路的路由尋徑目前仍鮮少受到討論，但是路由尋徑在網路架構當中依舊扮演著重要的角色。在本論文中在NDN中利用平行名稱查詢(Parallel Name Lookup, PNL)，將NDN以名稱前綴樹(Name Prefix Tree, NPT)的方式構築，讓通過路由器接口的資料封包做紀錄，紀錄各接口流量與接入檔案的類型，儲存成經歷記錄儲存至轉發資料庫當中，當接收到下一個興趣封包時，就能透過經歷記錄計算出各接口的流量與檔案類型，透過貝式分類法做演算選擇最佳的接口，藉以修正蟻群最佳化路由(Ant-colony Optimization Routing)的嗅探封包(Sniffing Packets)計算出的最短路徑之接口，來確保較短的傳送路徑而且避免蟻群最佳化中途點有目標資料的浪費情況。本論文以資料接口流量與資料類型做為路徑預判的決策，藉由此種分類決策增進尋徑效能並大幅減少資料搜尋時產生的興趣封包，同時降低路徑選擇的風險與帶來的衝擊。 Launched by the University of California at Los Angeles, the Named Data Networking (NDN) project aims to change the current Internet communication protocol between hosts. NDN uses data names instead of IP addresses for transmissions and develops a Content Store (CS) and a name-based routing mechanism to decrease the throughput and improve the data retrieval response time. In NDN, the network can identify the content and save it to the nearest router, thereby providing the optimal path from the client to the closest copy of the content. Therefore, efficient content delivery can be guaranteed by path and resources optimization. The Next Generation Network (NGN) has been a major research direction while the NDN is one of the most popular topics in NGN. Routing in NDN has been rarely discussed but it absolutely plays an important role in a new network architecture. In this paper, we use Parallel Name Lookup (PNL) to build the NDN by Name Prefix Tree (NPT). Our proposed scheme records the data packets received by the router, takes down the data flow and data types of each face, and save the information to the Forwarding Information Base (FIB). When the router receives the next interest packet, we can calculate the flow and the data type according to the record and choose the best face using a Bayesian classification. Through the sniffing packets of ant-colony optimization routing, our scheme can verify the face of the shortest path, ensure the shortest path and avoid the fast-path problem in ant forwarding mode. In this paper, data received by the face and the required data type is used for path prediction. By using such a classification method, our proposed scheme can improve the routing efficiency, significantly decrease the interest packets generated in data searching, and reduce the risks and impacts of path selection.
Appears in Collections:	[Graduate Institute & Department of Electrical Engineering] Thesis

Files in This Item:

File	Size	Format
index.html	0Kb	HTML	281	View/Open

Loading...