第四章模擬實驗與結果分析 4.1 模擬實驗的設計與假設本實驗使用 NS-2(Network Simulater ver. 2.29) 進行模擬, 我們另外使用 WiMAX_v2.3 module[13], 這是一個 NS-2 的 WiMAX 模組 由於 WiMAX_v2.3 module 為目前 NS-2 上有支援 WiMAX QoS 設計的模組, 因此我們將利用它來做為我們實驗的環境, 最後我們將在本章比較驗證實驗的結果 4.1.1. The MAC simulation architecture of IEEE 82.16 圖 4.1 The MAC simulation architecture of IEEE 82.16 [13]
由圖 4.1[13], 我們知道這個 WiMAX Module 已經具備基本 MAC layer 的所有功能, 如 connection admission control bandwidth request bandwidth allocation 及 scheduling, 只不過具備的功能都是以簡單的演算法運作 4.1.2. 系統模擬參數設定 表 4.1 Parameters Used in the MAC and Physical Layers[13] 由表 4.1[13] 我們知道, 系統在 MAC layer 以及 PHY layer 已經有一般化的設定, 因此使用 符合 82.16 standard 的 WiMAX module 將會是我們最好的選擇
4.1.3 Scenario 我們設計了一個範圍為 1m*1m 的 WiMAX 環境, 並且實驗 5 個 scenario, 首先第一個 scenario 為 8 個 node,8 個 node 依序建立不同類別的 traffic, 例如編號 1~4 對應 UGS rtps nrtps BE 連線 編號 5~8 也一樣對應到 UGS rtps nrtps BE 連線 Topology 如圖 4.2 所示, 模擬時間為 1 秒, 並各別傳送產生的 traffic 第二個 scenario 為 3 個 node, 依序建立不同類別的 traffic, 例如編號 1~5 對應 rtps rtps rtps nrtps BE 連線 編號 6~1 也一樣對應到 rtps rtps rtps nrtps BE 連線, 以此 類推 11~3 分別如上規則,topology 如圖 4.3 所示 第三個 scenario 為 3 個 node 建立 42 條連線 依序建立不同類別的 traffic, 編號 1~1 分 別為 rtps rtps rtps nrtps BE 連線循環, 而編號 11~3 則 2*rtPS 2*rtPS 2*rtPS nrtps BE 連線循環, 目的是為了加重系統的 loading,topology 如圖 4.4 所示 第四個 scenario 為 5 個 node, 並且這 5 個 node 個別建立 5 條 traffic 分別為 UGS ertps rtps nrtps BE 連線, 因此系統將有 5*5=25 條連線, 主要目的是為了了解 SS scheduling 的有效性,topology 如圖 4.5 所示 第五個 scenario 為 1 個 node 建立 6 條連線, 各個 node 各別建立 5*rtPS 連線以及 1*BE 連線, 目的是為了加重系統的 rtps 負載, 來觀察 BE 是否因此而 starvation,topology 如 圖 4.6 所示
圖 4.2 8 nodes in scenario 1 圖 4.3 3 nodes in scenario 2
圖 4.4 3 nodes in scenario 3 圖 4.5 5 nodes in scenario 4
圖 4.6 1 nodes in scenario 5 4.1.4. 實驗步驟與方法我們設計的實驗有一個實驗組及一個對照組, 實驗組為我們提出的方法, 稱之為 dynamic polling interval function (), 對照組為 WiMAX Standard 我們將計算系統的 total throughput 各連線的 average throughput rate 以及 fairness 4.1.5. 評估方法 我們使用 throughput average throughput rate fairness 做為我們評比的依據 4.2 模擬結果及分析我們從圖 4.7~11 的 average throughput rate 可以發現, 實驗組的數據均比對照組高, 表示其頻寬使用度較好, 尤其在此實驗中, 以 function 運作的 nrtps 以及 BE 連線可以得到相當不錯的 average throughput rate, 另外我們觀察到圖中的 rtps average throughput rate 皆在 WiMAX_v2.3 standard 之上, 因此我們得知使用 function 模式可以得到與 WiMAX_v2.3 standard 相近似的 rtps averager throughput rate, 這說明了雖然我們在追求
Fairness 上有優先考量 low priority connection 但是並沒有因此犧牲 rtps 的 QoS 等級要求, 在 averager throughput rate 上我們發現對於那些低優先權的連線, 我們至少可以提升 56Kbit/s 效能, 因此我們認為在 bandwidth request 以及 Fairness 的設計上的確有滿足原本設計的構想 我們再觀察圖 4.12~16 的 Fairness, 我們發現 Fairness 的值都.1 上下, 在 Fairness 上至少可以獲得 3% 以上的改善, 因此我們認為在我們設計下的 bandwidth allocation 及 scheduling 的確可以有效的反應出 Fairness 另外, 我們可以從圖 17~21 得出, 在 total throughput 至少可以獲得 2% 以上的改善, 因此我們認為我們所提出的 function 的確可以很有效的請求頻寬以及獲得頻寬 根據以上的實驗結果, 我們發現使用我們的 MAC co-function ( function) 的設計, 系統可得到較高的 throughput, 並且發現即使提高了 Fairness, 並不會降低 rtps 的 QoS 要求, 因此我們所提出的設計, 的確可以有效提升系統效能以及公平性
Kbits/s 25 2 15 1 Average Throughput BE_ nrtps_ rtps_ UGS_ BE_ nrtps_ rtps_ UGS_ rtps rtps std nrtps_ BE_ 5 nrtps BE 1 2 3 4 5 6 7 8 9 1 圖 4.7 average throughput in Scenario1 2 18 16 Average Throughput BE_ nrtps_ rtps_ rtps_std BE_ nrtps_ rtps_ 14 Kbit/s 12 1 8 6 rtps_ BE_ nrtps_ 4 nrtps_ BE_ 2.5 1 1.5 2 2.5 3 圖 4.8 average throughput in Scenario2
Kbit/s 18 16 14 12 1 8 6 4 2 Average Throughput BE_ nrtps_ rtps_ BE_ nrtps_ rtps_ rtps_ rtps_std nrtps_ BE_ nrtps_ BE_.3.6.9 1.2 1.5 1.8 2.1 圖 4.9 average throughput in Scenario3 25 2 15 Average Throughput BE_ nrtps_ rtps_ ertps_ UGS_ BE_ nrtps_ rtps_ ertps_ UGS_ rtps rtps Kbit/s 1 nrtps BE BE 5 nrtps 1 2 3 4 5 6 7 8 9 1 圖 4.1 average throughput in Scenario4
12 1 BE_ BE_ Average Throughput rtps_ rtps_ 8 rtps_ rtps_ Kbit/s 6 4 BE_ 2 BE_.3.6.9 1.2 1.5 1.8 2.1 2.4 2.7 3 圖 4.11 average throughput in Scenario5 1.9.8.7 Fairness.6 degree.5.4.3.2.1 1 2 3 4 5 6 7 8 9 1 sec 圖 4.12 Fairness in Scenario1
.5.45.4.35 Fairness degree.3.25.2.15.1.5.3.6.9 1.2 1.5 1.8 2.1 2.4 2.7 3 sec 圖 4.13 Fairness in Scenario2.5.45.4.35 Fairness degree.3.25.2.15.1.5.1.2.3.4.5.6.7.8.9 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2 2.1 2.2 sec 圖 4.14 Fairness in Scenario3
.6.5 Fairness.4 degree.3.2.1 1 2 3 4 5 6 7 8 9 1 sec 圖 4.15 Fairness in Scenario 4.45.4.35 Fairness.3 degree.25.2.15.1.5.3.6.9 1.2 1.5 1.8 2.1 2.4 2.7 3 sec 圖 4.16 Fairness in Scenario5
Total Throughput 12 1 throughput(kbyte) 8 6 4 4% 提升 2 1 2 3 4 5 6 7 8 9 1 圖 4.17 Total Throughput in Scenario1 Throughput(Kbytes) 16 14 12 1 8 6 Total Throughput 3% 提升 4 2.3.6.9 1.2 1.5 1.8 2.1 2.4 2.7 3 圖 4.18 Total Throughput in Scenario2
16 14 Total Throughput 12 Throughput(Kbytes) 1 8 6 2% 提升 4 2.3.6.9 1.2 1.5 1.8 2.1 圖 4.19 Total Throughput in Scenario3 2 18 16 Total Throughput Throughput(Kbytes) 14 12 1 8 6 2% 提升 4 2 1 2 3 4 5 6 7 8 9 1 圖 4.2 Total Throughput in Scenario4
25 2 Total Throughput Throughput(Kbytes) 15 1 3% 提升 5.3.6.9 1.2 1.5 1.8 2.1 2.4 2.7 3 圖 4.21 Total Throughput in Scenario5