| 摘要: | 行動機會網路是一種非典型的新型分散式網路，具非結構式的組織，且資料傳輸不需要仰賴傳統連線導向的路由機制，因此可以被應用在網路基礎脆弱、節點存在不確定、節點具高移動性或連線不易維持的網路環境，例如戰地網路、災難應急網路、耐延遲網路、車載網路及水下感測網路等。在行動機會網路環境下節點可透過在移動過程之中與其他節點相遇的機會，快速傳遞、交換及暫存訊息，以此方式逐漸將訊息轉送到目的端節點。然而，由於行動機會網路的應用發展相當著重於場域的真實性及網路的動態性，嚴苛的外在環境和受限的網路能力將會嚴重影響到訊息傳播技術的發展，要如何提出有效的研究方案，以協助行動機會網路服務和訊息傳輸技術之發展，是極負有價值的研究議題。 我們的研究發現在行動機會網路下訊息傳播的效能明顯會受到行動節點的移動特性之影響，節點的移動行為不僅具有群聚性，亦具有時間或空間的週期性，同時兩種週期性之間隱含有對稱性等特徵。若能在行動機會網路的研究中導入合適的節點移動模型，並結合時間與空間週期性的移動特徵，這將有利於發展高效性的訊息傳輸技術與訊息管理策略。因此，本計劃的目標擬研發一系列適用於行動機會網路環境的訊息傳遞機制以及儲存資源分配的策略；研究內容主要包括以下五個部分，依序進行：第一部分，分析時間維度上週期序列的變異，提出一套具動態週期性感知能力的訊息傳遞方法；第二部分，以機率性圖論模型來表示週期性相遇的不確定性，據以提出合適的訊息傳遞方法；第三部分，觀察節點相遇週期在時間和空間維度上的對稱性，據以提出一套結合時間與空間週期性的訊息傳遞方法；第四部分，納入巡迴節點的編組，設計一套基於協作式巡迴節點的訊息傳遞方法；第五部分，研發一套基於賽局理論的節點儲存資源分配之機制，評估對系統整體效能優化之成效。最後，本計劃執行的成果將具體提出行動機會網路中具週期性移動行為之參考模型，完成上述的新型訊息傳輸及管理機制之設計及其效能的量測和分析。 ;Sustainable resource provision amongst upgraded mobile nodes in a network enables mobile nodes for collaborating in network construction, message delivery and resource sharing, definitely introducing a non-classical mobile opportunistic network. Mobile opportunistic networking is a modern distributed network architecture: particularly, a system is based on unstructured organizations and supports transient communications without resorting to traditional connection-oriented routing paradigms. Mobile opportunistic networking technologies are believed to be suitable in use for many pragmatic application domains, e.g., battlefield communications, disaster response networks, delay/disruptive networks, vehicular ad hoc networks, and underwater acoustic networks. This sort of practical networks is characterized by fragile network topologies, uncertainty of node presence, high node mobility, and sporadic connections. Message dissemination in a network must depend on relay nodes that repeatedly forward messages to neighbor nodes that are encountered during their movements. Owing to the aspects of environmental affects in reality and network dynamics, however, it is difficult to guarantee the applicability and performance of message dissemination mechanisms. In response to critical circumstances above, a promising solution is essential for the design and development of mobile opportunistic message dissemination technologies, thereby being a vital research dimension. Our study in the literature and many practical traces finds that node mobility patterns impose a significant effect on message dissemination in mobile opportunistic networks. The behavior of node mobility, empirically, involves flash crowding, clustering and periodicity properties at spatial and/or temporal standpoints, as well as implicit symmetry between spatial and temporal relationships. Prospectively integrating appropriate node mobility models with temporal and spatial periodicities in mobile opportunistic networks would be much beneficial for message dissemination and service designs in highly dynamic and unreliable network environments. Therefore, this research project aims to develop new message dissemination and management techniques for mobile opportunistic networking. This project proposal comprises five parts, as follows. First, we will analyze the variance of temporal periodicity and then design a dynamic periodicity-aware message delivery scheme. Second, we will use a probabilistic space-time graph model to develop a periodic message delivery scheme. Third, we will design a message delivery scheme based on spatial-temporal periodic symmetry. Fourth, we will account for the ferry-aided effects and then design a cooperative-ferry-based delivery scheme. Finally, we will adopt the game theory for performance optimization on storage utilization among all nodes in a network system. In accordance with all above, our study will achieve a new node mobility reference with spatial and temporal periodicity, and a dedicated suite of message dissemination and management mechanisms in mobile opportunistic network environments. |
