Design of Privacy Protection Schemes for Mobile Adhoc Networks using Rough Set Theory

Abstract

MANET is a self-con guring, decentralized and infrastructure-less mobile wireless network, where autonomous mobile nodes (such as laptops, smartphones, sensors, etc.) communicate over the wireless channels. Thus, MANETs are suitable for supporting decentralized and mo- bile applications in the areas, like healthcare, military, commercial, education and many others. However, MANETs consist of limited transmission range, limited device capabilities, unreliable wireless links, dynamic changes in network topology, lack of security, etc., which introduce sev- eral vulnerabilities that must be dealt with to achieve the success of MANETs. A vulnerability is a weakness, that is exploited by the attackers to introduce privacy breaches during route es- tablishment and data transfer stages. In route establishment stage, the nodes have to disclose some of the routing information such as their identity, location, etc., which can be obtained by an attacker at the malicious intermediate nodes, thus resulting in privacy breach. Due to wireless nature of links between the mobile nodes, the attackers can identify the communicating nodes just by overhearing and tracing the transmitted data packets from sender to receiver, thus breaching privacy during the data transfer stage. Thus, lack of strong privacy preserving solutions may lead to highly unacceptable results, such as real-time tracking of location move- ments, the disclosure of sensitive data, etc. In this thesis, we propose solutions to preserve location privacy and data privacy in MANETs, which are paramount to ensure acceptance of MANET applications. We have applied Rough Set Theory (RST) concepts, as it is efficiently used for classi cation of nodes, and generating privacy policies with minimum overheads by eliminating redundant information. First, we design and develop, a Location Privacy Preservation (LPP) protocol which estab- lishes an untraceable route between communicating nodes while preserving location privacy. The sender initiates the establishment of a route to a receiver through trusted nodes, where RST de nes the trust attributes such as resource availability, node reliability and node history of 1-hop neighbor nodes. The trust value (or trustworthiness) of a node is determined based on their trust attribute values. The route between sender and receiver nodes is established through trusted nodes only, and these trusted nodes act as the temporary sender for their next hop, till the receiver is reached. To ensures that the route is established with designated trusted node, and to check for any suspected situation, challenge-response messages are exchanged at each hop. The proposed LPP protocol is evaluated through simulation and also compared with earlier works. We discuss the performance analysis of the LPP protocol. We provide a formal veri cation model to shows the validity of the LPP protocol using ProVerif tool (an automatic formal protocol veri cation tool), which is utilized to formalize the functions of LPP protocol using ProVerif's calculus. The data transmitted may contain sensitive information, and undesired disclosure of in- formation can lead to the launching of various attacks, thus breaching the data privacy. For this, we have proposed a Data Privacy Preservation (DPP) scheme based on data anonymity approach, where RST concepts are applied to determine the level of data anonymity during the data transfer. Data packets are enclosed within capsules that can be opened only by the designated nodes, thus preventing the undesired leakage of the data. On the suspected situa- tions, challenge-response messages are exchanged to check data privacy violations by next hop trusted node. The route between sender and receiver is changed dynamically at each hop. The proposed DPP scheme is evaluated through simulation, and compared with some of the earlier works. The performance analysis of DPP scheme is discussed. Also, DPP scheme is tested by considering different case studies in a MANET deployed for the stock market application. Due to the development in the technology and realistic result expectation of the user, it is required to develop more practical privacy preserving solutions by collecting the context information. In this connection, we have extended the LPP protocol to a Context based Loca- tion Privacy Maintenance (CLPM) scheme, which takes context information related to nodes en route to maintain location privacy during the data transfer. RST concepts are applied to determine the privacy maintenance level of next hop trusted node, and then based on the pri- vacy maintenance level, failure to preserve location privacy are identi fied. The challenge and response messages are exchanged at each hop to ensure that trusted node maintains location privacy. If location privacy violation is detected, then the route is locally repaired. Otherwise, if there is no location privacy violation, then data is transferred through the same trusted node. The performance of CLPM scheme is evaluated through simulation, and compared with the earlier works. The performance analysis of CLPM scheme is carried out. Healthcare is currently, one of the most attractive application areas in the Internet of Things (IoT), which includes many benefi ts such as real-time patient monitoring, elderly care, and much more. However, privacy in IoT healthcare remains the most challenging obstacle, which requires dynamic privacy protection solutions for preserving the privacy of patients, doctors, etc. Thus, proposed privacy preserving protocol and schemes are applied to preserve location privacy and data privacy in the IoT healthcare application. Then, some of the healthcare transactions are illustrated to test the working of proposed privacy preserving protocol and schemes. In summary, in this work, we have designed: 1) a location privacy preservation protocol which establishes an untraceable route between sender and receiver, and preserves location pri- vacy; 2) a data privacy preservation scheme based on data anonymity approach, where concepts of RST are applied to hide the sensitive data during data transfer; 3) a context based location privacy maintenance scheme, which takes context information of nodes en route to maintain location privacy; and 4) an IoT healthcare application is considered to test the designed pro- tocol and schemes for preserving location privacy and data privacy. Simulations are performed in different MANET environments to test the proposed protocol and schemes. The simulation and analytical results obtained show the importance of the approaches and the efficiency of the proposed protocol and schemes, which could be implemented in the existing applications. Based on obtained solutions, the system can be applied to any network by adapting the nature of communications and security challenges of that network.