الفهرس 
ContentsOnly 14 pages are availabe for public view
 |  | from | 105 |  |  |
| | | from | 105 | | |
Abstract
Mobile ad hoc network (MANET) is a new and evolving area of interests, which used in many different applications. Black hole attack is a Denial of Service (DoS) attack; it works by drawdown packets in the network to malicious node and then drops, alters the content of the packets, or even passes the packets to another malicious node. Another effective attack is gray-hole attack, MANET works under an assumption that all nodes in the network are collaborating to forward packets, which is not true as there are selfish nodes that refuse to forward the packets to reserve its energy and other resources, also there are attack nodes, which DROP packet to harm the network. One of the important parts in utilizing and deploying the MANET network is securing it. Achieving a secure MANET helps this kind of network to achieve its full potential, which is to be used not only in military and crises situation applications but also in a commercial way. A certain level of security can accomplish by using the existing security solution. However, because of the nature of the MANET, it has its own vulnerabilities coupled with the normal vulnerabilities of the wireless networks. Therefore, these solutions cannot provide a sufficient security level. Intrusion detection systems with the traditional security solutions can accomplish a sufficient security level. In this thesis,(i) A proposed intrusion detection system (IDS) introduced against the black-hole attack, where an adaptive neuro-fuzzy inference system (ANFIS) used to automate the process of producing a fuzzy system and then optimizes this system using the genetic algorithm (GA); (ii) A comparison between normal estimated fuzzy interface system and the optimized fuzzy interface system is done to evaluate the performance of the two systems ; (iii) An evaluation of the acknowledgement (ACK) based IDS techniques against selective forwarding gray-hole attack is presented to evaluate the ability of the ACK based techniques to overcome a major disability in watchdog technique which detecting malicious misbehaviors with the presence of partial dropping (50%). For detecting and preventing black hole attack, it is proven that the optimized proposed system was improved in an average of 5% from the estimated system in low speed and improved to an average of 9% from the estimated system in high speed in the packet delivery ratio. However, an increase of an average of 0.1% in the routing overhead in low speed is noticed with an increase of an average of 6% in high speed. For gray-hole attack, it is proven that the ACK based techniques can actually overcome this disability but only with low speed but with high speed and the presence of collaborative attacks the ACK, based techniques have low performance. Based on simulation results, we have proposed a modified fuzzy based IDS algorithm to provide optimal network performance in the presence of black hole attack. These results were verified using Packet deliver ratio and routing overhead performance matrices.