A Secure Protocol against Selfish and Pollution Attacker Misbehavior in Clustered WSNs

Wireless sensor networks (WSNs) have been widely used for applications in numerous fields. One of the main challenges is the limited energy resources when designing secure routing in such networks. Hierarchical organization of nodes in the network can make efficient use of their resources. In this case, a subset of nodes, the cluster heads (CHs), is entrusted with transmitting messages from cluster nodes to the base station (BS). However, the existence of selfish or pollution attacker nodes in the network causes data transmission failure and damages the network availability and integrity. Mainly, when critical nodes like CH nodes misbehave by refusing to forward data to the BS, by modifying data in transit or by injecting polluted data, the whole network becomes defective. This paper presents a secure protocol against selfish and pollution attacker misbehavior in clustered WSNs, known as (SSP). It aims to thwart both selfish and pollution attacker misbehaviors, the former being a form of a Denial of Service (DoS) attack. In addition, it maintains a level of confidentiality against eavesdroppers. Based on a random linear network coding (NC) technique, the protocol uses pre-loaded matrices within sensor nodes to conceive a larger number of new packets from a set of initial data packets, thus creating data redundancy. Then, it transmits them through separate paths to the BS. Furthermore, it detects misbehaving nodes among CHs and executes a punishment mechanism using a control counter. The security analysis and simulation results demonstrate that the proposed solution is not only capable of preventing and detecting DoS attacks as well as pollution attacks, but can also maintain scalable and stable routing for large networks. The protocol means 100% of messages are successfully recovered and received at the BS when the percentage of lost packets is around 20%. Moreover, when the number of misbehaving nodes executing pollution attacks reaches a certain threshold, SSP scores a reception rate of correctly reconstructed messages equal to 100%. If the SSP protocol is not applied, the rate of reception of correctly reconstructed messages is reduced by 90% at the same case.

Wireless Mesh Network Security, Architecture, and Protocols

Due to suppression of central administration in WMN, network functioning like network controls, management, routing, switching, packet forwarding etc. are distributed among nodes, either collectively or individually. So, cooperation among nodes is highly solicited. However, there may exist node's malicious activities because of its open characteristics and limited available battery power. The nodes may misbehave by refusing to provide service or dropping down the packets because of its selfishness and malicious activity. The identification of misbehaving nodes and prevention from them can be one of the biggest challenges. Hence, the prime target of the chapter is to provide an overview of existing intrusion detection and prevention approaches, and secure routing or framework that can recognize and prevent from the malicious activities. The digital signature-based IDS to offer secure acknowledgment and an authentication mechanism has also been discussed. The expectation is the digital signature-based IDS will overcome the weakness of existing IDS.

Feature Reduction in MANET using Machine Learning Language

Mobile ad-hoc network (MANET) is an infrastructure-less network. Therefore, MANET involves a selection of exact security schemes to notice the false entrance of the mischievous nodes. Along these lines, we require solid instrument to identify these pernicious nodes and to arrange ordinary and irregular nodes based on the conduct or performance of nodes. Machine learning system nowadays used to built a best IDS for recognizing exception or misbehaving nodes rapidly and precisely give grouping by watching conduct of those nodes in the system. In MANET system, numbers of parameters are taken for analysation. It makes the IDS system complex. To avoid this complexity many techniques are derived for feature reduction. In this proposed work, we are testing how feature reduction can be done using Python machine learning program.

Dynamic Trust Management for Community Based Mobile Grid Application

Mobile Grid is the inter-networking of heterogeneous physical as well as virtualdevices. Each device transfer and share the information with each other. Trust management plays a significant role in network based applications for information collection, data mining, qualified services with context-awareness, upgraded client protection and data security. It assists individuals with beating impression of vulnerability, threat and participates in client acknowledgment to utilization on grid services and applications. In this paper a unique trust management protocol is proposed for network based mobile grid application to manage misbehaving nodes whose status or performance may change dynamically. Trust plays an important role for handling the security in the community based system. Trust management provides facilitate to identify malfunctions and also make legitimate collaboration and enhance the user privacy and information security.