隨意式網路中監聽節點之偵測與隔離

Abstract

隨意式網路中有些攻擊是惡意節點必須處在監聽模式並與其它節點執行相同繞路協定才能進行的，針對此類攻擊，我們提出了一個偵測監聽與隔離的機制。利用在AODV 中增加一個新的控制封包來偵測相鄰節點是否有做監聽的動作，並且透過隔離機制將監聽節點隔離在AODV 網路之外，使其無法進行攻擊。本文使用Ns-2 模擬軟體來評估我們設計的偵測與隔離機制，結果顯示我們的機制不但能有效偵測及隔離監聽節點，同時對網路的負擔也不大。

In the AD HOC, some attacks which could be achieved must be under sniffing mode and under the same routing protocol. Aimed at these attacks, we propose a detect sniffing and isolate mechanism. We add a new control packet to detect neighbor nodes are sniffing or not, and use isolate mechanism to isolate the sniffing node out of AODV network, then the sniffing node can not attack any more. In this paper, we use Ns-2 simulator to estimate our detection and isolation mechanism. From the simulation result, we can see that our detect mechanism can detect and isolate sniffing nodes effectively and will not give too much load to network.