Artificial immune system based security algorithm for mobile ad hoc networks

摘要

Securing Mobile Ad hoc Networks (MANET) that are a collection of mobile, decentralized, and self-organized nodes is a challenging task. The most fundamental aspect of a MANET is its lack of infrastructure, and most design issues and challenges stem from this characteristic. The lack of a centralized control mechanism brings added difficulty in fault detection and correction. The dynamically changing nature of mobile nodes causes the formation of an unpredictable topology. This varying topology causes frequent traffic routing changes, network partitioning and packet losses. The various attacks that can be carried out on MANETs challenge the security capabilities of the mobile wireless network in which nodes can join, leave and move dynamically. The Human Immune System (HIS) provides a foundation upon which Artificial Immune algorithms are based. The algorithms can be used to secure both host-based and network-based systems. However, it is not only important to utilize the HIS during the development of Artificial Immune System (AIS) based algorithms as much as it is important to introduce an algorithm with high performance. Therefore, creating a balance between utilizing HIS and AIS-based intrusion detection algorithms is a crucial issue that is important to investigate. The immune system is a key to the defence of a host against foreign objects or pathogens. Proper functioning of the immune system is necessary to maintain host homeostasis. The cells that play a fundamental role in this defence process are known as Dendritic Cells (DC). The AIS based Dendritic Cell Algorithm is widely known for its large number of applications and well established in the literature. The dynamic, distributed topology of a MANET provides many challenges, including decentralized infrastructure wherein each node can act as a host, router and relay for traffic. MANETs are a suitable solution for distributed regional, military and emergency networks. MANETs do not utilize fixed infrastructure except where a connection to a carrier network is required, and MANET nodes provide the transmission capability to receive, transmit and route traffic from a sender node to the destination node. In the HIS, cells can distinguish between a range of issues including foreign body attacks as well as cellular senescence. The primary purpose of this research is to improve the security of MANET using the AIS framework. This research presents a new defence approach using AIS which mimics the strategy of the HIS combined with Danger Theory. The proposed framework is known as the Artificial Immune System based Security Algorithm (AISBA). This research also modelled participating nodes as a DC and proposed various signals to indicate the MANET communications state. Two trust models were introduced based on AIS signals and effective communication. The trust models proposed in this research helped to distinguish between a “good node” as well as a “selfish node”. A new MANET security attack was identified titled the Packet Storage Time attack wherein the attacker node modifies its queue time to make the packets stay longer than necessary and then circulates stale packets in the network. This attack is detected using the proposed AISBA. This research, performed extensive simulations with results to support the effectiveness of the proposed framework, and statistical analysis was done which showed the false positive and false negative probability falls below 5%. Finally, two variations of the AISBA were proposed and investigated, including the Grudger based Artificial Immune System Algorithm - to stimulate selfish nodes to cooperate for the benefit of the MANET and Pain reduction based Artificial Immune System Algorithm - to model Pain analogous to HIS.