行動隨意網路下之調節式邊緣認知廣播策略

Abstract

在行動隨意網路中，對於許多應用而言廣播是一項必要功能，例如，路由搜尋、位址解析等應用。傳統的廣播方法因為重覆廣播、競爭及碰撞等問題而浪費很多網路資源，所以已經有很多方法被提出來解決這些問題。但是很少有方法可以同時適用稀疏式和密集式網路。在本篇論文中，我們提出一個調節式邊緣認知的廣播策略 (ABB)，ABB只需要1-hop節點的資訊，因此可以減少控制訊息過載。我們提出了兩個改進方法：第一個方法稱為稀疏模式改進，在稀疏式網路中這個改進藉由減少取消的範圍來增加到達率。第二個方法稱為密集模式改進，在密集式網路中藉由這個改進可以減少重覆廣播的範圍來獲得高的重覆廣播節省率。評估結果顯示，我們所提出的ABB方法在所有網路環境下都可以有較高的到達率，所以無論是在密集或稀疏式行動隨意網路之下，我們的方法均可適用。就平均而言，在到達率方面，ABB可以達到和ACB一樣的水準，然而可以比DFCN好33%。在平均廣播封包的延遲時間方面，ABB比ACB好33%，而且比DFCN好7%。另外，在重覆廣播節省率方面，ABB比ACB好16%，但是比DFCN差35%。然而DFCN有比較低的到達率，而對於評估廣播方法而言，到達率是一個比較重要的參數。

In mobile ad hoc networks (MANETs), broadcast is an essential function for many applications, such as route discovery and address resolution. Since conventional broadcast in MANETs wastes a lot of resources due to redundant rebroadcasts, contention and collision, several existing broadcast protocols were proposed to resolve these problems. However, few of them are suitable to sparse networks as well as dense networks. In this thesis, we propose an Adaptive Border-aware Broadcast (ABB) that requires only 1-hop neighbor information so as to reduce control overhead. Two enhancements are proposed. The first enhancement, called sparse mode enhancement, can increase the reachability by decreasing the “cancellation range” in sparse networks. The second enhancement, called dense mode enhancement, can reduce the number of rebroadcast nodes in order to have high saved rebroadcast by decreasing the “rebroadcast range” in dense networks. Simulation results show that the proposed ABB can achieve high reachability in all network environments. Therefore, our ABB is suitable for both dense and sparse MANETs. Averagely, ABB is comparable to ACB and is 33% better than DFCN in terms of reachability. For average latency, ABB is 33% better than ACB and 7% better than DFCN. In addition, for saved rebroadcast, ABB is 16% better than ACB but 35% worse than DFCN. However, DFCN has poor reachability, which is a more important metric for evaluating broadcast schemes.