對無線移動隨機式網路中資料可獲取度之策略探討

Abstract

無線移動隨機式網路在民間和軍事環境中都有各式各樣的應用，所以越來越受到重視。然而，這種網路因為本身具備的一些特有性質而對研究上造成許多挑戰。我們主要的目的是設計出一種資料傳送的協定，這種協定能讓當網路中的節點消失或點對點的連結斷掉時，整個網路還是連通使資料傳遞能受到比較小的影響。另外，在同儕計算的網路中，資料可獲取性也變成很重要的議題；事實上，目前已經有大量的探討在於同儕計算中資料存取方面。 在這篇論文中，我們首先介紹一些基本的問題，這些問題即使是處於相當簡化的局面下，也因為應用在無線移動隨機式網路裡的同儕計算的關係而變的非常棘手。然後，我們設計了一個局部路徑修復的方法來提高整個網路的資料可獲取性。此方法是用分散的方式來執行，也就是只有每個節點的局部資料而不是整個網路的資料被傳遞。最後，為了對不同種的網路性質（隨機圖形、無尺度網路、移動隨機式網路）有更完整的瞭解，我們跑了一些實驗模擬並對結果作詳細分析和討論。

Wireless ad-hoc networks are drawing significant attention because of the various potential applications in civilian and military environments. The inherent characteristics of these networks, e.g., lack of physical infrastructure and unreliable communication links, however, pose many challenges. A prime goal is to design data dissemination protocols that are resilient to abrupt malfunction of links and nodes. In Addition, data availability appears an attractive issue in peer-to-peer (P2P) overlaying networks; in fact, tremendous studies in application-level P2P data retrieval have been presented. In this thesis, we illustrate the intractability of several fundamental problems related to peer-to-peer data access in wireless ad-hoc networks, even under a rather simplified scenario. We then design a method used for localized path repair, which could improve the overall data accessibility; moreover, our approach is performed in a decentralized manner and relays on local knowledge of each node due to the ad-hoc nature. Finally, we do several simulation and analysis to acquire more complete understanding to the characteristics of different classes of networks: correlated random networks, scale-free networks and wireless ad-hoc networks.