| 摘要: | 由於節點之間的連線在耐延遲網路(DTNs)下存在間歇性連線的問題，因此在這樣的環境下設計訊息傳遞機制極具挑戰性。間歇性的連線使得節點很難找到一條來源端到目的端的點對點傳送路徑。因此，在DTNs的環境中，訊息的傳遞必須透過儲存、攜帶與轉送的方式。
在真實情況下，人類的移動行為並非隨機的。相反的其受節點之間的社群關係所影響。然而因為網路拓譜的動態變化，使得要建立能廣泛適用的節點社群關係較為困難。由於上述動機，本論文藉由節點之間的社群關係所造成的節點群聚現象，設計一套新穎的動態叢集感知建構之訊息傳遞機制，簡稱MDCA。其是基於節點的社群關係使節點在網路中所造成的群聚現象。如果訊息可以被傳送到各個叢集，則訊息傳達率將可以被改善
為了達到此目的，MDCA被細分為七個程序: (1) 面積估測 (2) 叢集決策 (3) Quality 值更新 (4) 訊息密度決策 (5) 根據節點的數個度量將節點排名 (6) 節點進入/離開叢集決策 (7) 訊息傳遞。主要做法如下。首先，在程序(1)節點會計算叢集節點密度的期望值。接著在程序(2)節點會判斷是否正位於叢集內。最後透過程序(5)和(6)，節點篩選出適合的中繼節點將訊息留在叢集內或散播到各個叢集。此外，MDCA透過訊息密度決策來控制訊息的副本數量，也就是程序(4)。最後，本論文在不同的移動模型上進行大量模擬，包含Random Waypoint (RWP)、Time-Variant Community Mobility model (TVCM)以及真實軌跡檔Infocm06。模擬結果顯示，MDCA在具有人類移動行為的移動模型上有較好的訊息傳達率。
;Designing a message forwarding scheme in delay-tolerant networks (DTNs) is a challenging problem due to intermittent connectivity between nodes. The problem of intermittent connectivity makes a node difficult to find an end-to-end path for any source-destination pair in a network. Therefore, the store-carry-and-forward messaging method is used in DTNs.
Movement behavior of humans in real life is not random. A societal relationship exists among nodes. However, dynamic changes of network topology make it difficult to define the general relationship between nodes in a network. This problem motivates the study of this thesis to design a novel message forwarding scheme with dynamic cluster awareness (MDCA), which exploits the nodes’ aggregation phenomenon caused by an implicit relationship among nodes in a network. Because the aggregation phenomenon of nodes will create some clusters in a network, the delivery probability can be improved if messages can be distributed to each cluster.
To achieve this goal, the MDCA design includes seven functional processes: (1) area estimation, (2) cluster decision, (3) quality value update, (4) determining message density, (5) ranking nodes’ metrics, (6) nodes moving in/out of a cluster, and (7) message transmission. Thus, the main idea of MDCA is as follows. First, the node calculates the expected density of nodes in process 1. Secondly, the node determines whether it is in a cluster or not in process 2. Lastly, the node chooses the appropriate relay nodes to carry messages in a cluster through processes 5 and 6. In addition, the MDCA controls the quantity of message copies by determining the message density, i.e., process 4. Furthermore, this study conducts many simulations with various mobility models, including random waypoint (RWP), time-variant community mobility model (TVCM) as well as real trace Infocm06. Performance results show that the MDCA has better delivery probability in the mobility models related to human behavior. |
