標題: 網路錯誤與網路威脅診斷之蟻群優化研究
Diagnosing Network Faults and Network Threats Using Ant Colony Optimization
作者: 陳夏祥
黃世昆
Chen, Hsia-Hsiang
Huang, Shih-Kun
資訊科學與工程研究所
關鍵字: 統合式系統框架;網路錯誤點;軟體弱點;錯誤定位;網路服務品質錯誤點;蟻群演算法;拒絕存取服務攻擊;網路服務品質攻擊;分散式拒絕存取服務攻擊;unified framework;network fault;software vulnerability;fault localization;QoS fault;ant agent system;DoS attack;QoS attack;DDoS attack
公開日期: 2016
摘要: 本篇論文針對網路錯誤與網路威脅,提出一個統合式系統框架的解決方案,此系統框架可以同時解決這兩種問題。而且問題的性質可分為兩大類,第一類為無人為意圖情況下所發生的網路錯誤問題,另一類為軟體潛在弱點被人為蓄意攻擊的問題。 因此,我們提出了兩種方法去解決此問題,並且可以同時對抗拒絕存取服務攻擊 (denial of service attack),網路服務品質攻擊 (quality of service attack) 和網路服務品質錯誤點 (quality of service fault) 的情況。而且對於分散式拒絕存取服務 (distributed denial of service) 透過此系統框架也可以有效偵測異常流量與確認攻擊路徑。此系統框架可經由ant colony system-based的方法快速地過濾異常封包和確認攻擊者來源,達到降低網路威脅的傷害和有效防範的目的。 此外,此系統框架針對在軟體定義網路 (software defined networking) 環境下,也運用spectrum-based軟體錯誤定位方法,此方法可以很精確地診斷出網路環境中的錯誤點,以及多重QoS fault的問題。最後,實驗結果證明所提出的方法能夠有效率,並且準確的找出攻擊來源和錯誤點。
In our work, we propose a unified framework to combine security faults and threats into a generalized behavior. That is, one is the unintended activity to trigger network faults and the other is the manual attack to trigger potential software vulnerability. We therefore propose two methods for dealing with spectrum-based software fault localization method for diagnosing network faults and multiple QoS fault cases. We also propose a network threat fast filtering and identification system by ant agent system to defend against DoS attack, QoS attack and QoS fault cases and an ant colony system for distributed detection and identification of DDoS attacks. As a result, the unified model has better performance than other methods in efficiency and effectiveness from our experiments.
URI: http://etd.lib.nctu.edu.tw/cdrfb3/record/nctu/#GT079555853
http://hdl.handle.net/11536/143366
Appears in Collections:Thesis