一般來說，運用鏈路負載平衡設備可以降低企業網路營運成本，且能強化其網路傳輸效能。有多種演算法可供鏈路負載平衡設備應用，以控管企業網路中多條網際網路鏈路的流量負載。每一種演算法均能應用在不同的量測距離和流量配送機制上，以不同的量測與配送成本來處理流量負載平衡。就鏈路負載平衡演算法而言，哪一種是最佳的距離量測與流量配送模式組合，以最低的成本獲得最佳的傳輸效能，這個議題值得進一步討論。
本研究嘗試藉由鏈路負載平衡演算法的探討，找出一種更有效率的距離量測和流量配送機制組合，期可獲得較佳的效能。首先，本研究嘗試瞭解在不同的距離量測和流量配送機制中，鏈路負載平衡機制的運作情形。然後依據量測距離與流量配送機制的運作情形，在研究中進行量測距離與流量配送機制的成本分析探討。
在本研究中建置一個鏈路負載平衡仿真環境，就不同流量量測與配送機制搭配不同的演算法進行仿真。運用最常使用的鏈路負載平衡演算法：輪送演算法(RR)和權重最少流量優先演算法(WLTF)，搭配不同流量量測與配送機制進行仿真。經檢測仿真所得的資料，發現輪送演算法和權重最少流量優先演算法均可由最低的量測與配送成本，獲得較佳的鏈路負載平衡效能。

Generally speaking, an enterprise network operation cost could be reduced and its performance would be enhanced with a link load balance device. However, there are many algorithms can be used by a link load balance device to handle traffic loading over multiple internet links within an enterprise network. Each algorithm can apply different measurement distances and traffic dispatching schemes to process traffic load balance with different measurement and dispatching cost. Which are the best combinations of measurement distance and traffic dispatching schemes for link load balance algorithms to receive better performance with lower cost? It is an issue needs discuss further.
This study tries to discuss which combination of measurement distance and traffic dispatching schemes is more efficient for link load balance algorithms to receive better performance. First, this study tries to understand the operation of a link load balance algorithm with different measurement distances and traffic dispatching schemes. According to the operations of measurement distances and traffic dispatching schemes, a cost analysis of measurement distances and traffic dispatching schemes is discussed in this study.
In this study, link load balance emulation environment is implemented to emulate link load balance processes which are handled by different algorithms with different traffic measurement and dispatching schemes. Several most used link load balance algorithms, round robin (RR) and Weighted Least Traffic First (WLTF), with different traffic measurement and dispatching schemes are emulated. By examining the emulated data, an interested summary can be found: the RR and WLTF algorithms can receive better link load balance performance with the lowest measurement and dispatching cost.

第一章 緒論.....................................................................................................1
1.1 研究背景與動機.................................................................................1
1.2 研究目的.............................................................................................1
1.3 研究方法與步驟.................................................................................3
1.4 研究範圍與限制.................................................................................5
1.5 論文架構.............................................................................................5
第二章 企業鏈路負載平衡機制網路環境之探討.........................................7
2.1 負載平衡機制.....................................................................................7
2.1.1 避免企業單一網際網路鏈路故障...........................................8
2.1.2 降低企業營運成本...................................................................9
2.1.3 平衡企業網路聯外鏈路之流量.............................................10
2.1.4 提升企業網路聯外鏈路服務品質.........................................11
2.1.5 建構企業所需網路安全環境.................................................12
2.1.6 可即時掌握的企業網路資訊.................................................13
2.2 應用DNS機制平衡企業網路聯外鏈路負載.................................14
2.3 企業網路網際網路鏈路流量負載平衡的方向...............................15
第三章 企業鏈路負載平衡機制量測與配送作業.......................................17
3.1 企業鏈路負載平衡機制運作模式...................................................17
3.1.1 企業鏈路負載平衡機制作業環境.........................................17
3.1.2 企業鏈路負載平衡機制運作流程.........................................18
3.2 企業網路負載平衡機制鏈路量測...................................................19
-iv3.2.1
鏈路負載平衡機制量測距離.................................................19
3.2.2 鏈路負載平衡機制量測頻率.................................................21
3.3 企業網路負載平衡機制流量配送...................................................24
3.3.1 負載平衡機制流量配送模式.................................................24
3.3.2 負載平衡機制流量配送模式成效分析.................................26
3.4 企業網路負載平衡機制作業演算法...............................................29
3.4.1 靜態配送演算法.....................................................................29
3.4.2 動態配送演算法.....................................................................32
第四章 仿真結果與分析...............................................................................36
4.1 仿真環境與想定說明.......................................................................36
4.2 線路流量距離量測與配送構面分析...............................................37
4.2.1輪送演算法量測距離與流量配送模式成效分析..................38
4.2.2 WLTF演算法距離量測與分析..............................................42
4.3 流量配送模式分析...........................................................................45
4.3.1 以「網際網路位址」與「連線」在近端之比較.................46
4.3.2 LTF演算法與WLTF演算法流量配送模式之分析.............47
第五章 結論與未來研究方向.......................................................................49
5.1 結論...................................................................................................49
5.2 未來研究方向...................................................................................49
參考文獻.........................................................................................................50

中文部分
[1] 中華電信股份有限公司 (民九五) http://www.cht.com.tw/BusinessCat.php?CatID=69。
[2] 彭祖乙（民九四），「多路徑負載平衡演算法：特徵、效能分析及改良」，國立交通大學資訊管理研究所博士論文。
[3] 楊憲其(民九一)，「在IPv6環境下Multihoming設計之研究」，國立中正大學資訊工程研究所碩士論文。
[4] 德恩資訊產品介紹—負載平衡器，http://www.deansoft.com.tw/C_balancer01.htm
英文部分
[5] D. Applegate and E. Cohen, “Making Intra-Domain Routing Robust to Changing and Uncertain Traffic Demands: Understanding Fundamental Tradeoffs,” in Proc. ACM SIGCOMM, 2003.
[6] Feldmann, A. Greenberg, C. Lund, N. Reingold, J. Rexford, and F. True, “Deriving Traffic demands for operational IP networks: Methodology and experience,” IEEE/ACM Transactions on Networking, vol. 9, no. 3, pp. 265–279, 2001.
[7] Gan, D. and Srisuresh, P. [1998], “Load Sharing using IP Network Address Translation (LSNAT)”, RFC2391.
[8] He, J. [2000], “An Architecture for Wide Area Network Load Balancing” Proceedings of the 2000 IEEE international Conference on Communications, vol.2, pp. 1169-1173, USA, New Orleans.
[9] H. R¨acke, “Minimizing congestion in general networks,” in Proc. 43rd IEEE Symp. Foundations of Computer Science, 2002., pp. 43–52.
[10] Quoitin, S. Uhlig, C. Pelsser, L. Swinnen, and O. Bonaventure, .Interdomain Traffic engineering with BGP,. IEEE Communications Magazine, pp. 122.128, May 2003.
[11] Jin, S., Wu, Q. and Zhang, W. [2000], “Scaling Internet services by LinuxDirector”, The Fourth International Conference/Exhibition on High Performance Computing in the Asia-Pacific Region, vol.1, pp. 176-183, China, Beijing.