Efficient Data Forwarding Protocols In Non-cooperative Mobile Social Networks

Mobile Social Networks(MSNs)are emerged as a novel wireless communication paradigm in which the social characteristics and behaviors of mobile nodes(i.e.,users and their carried devices)are exploited to promote their interactions.In MSNs,the performance of data delivery heavily relies on the cooperation level of nodes.However,some nodes may exhibit selfish behav-ior and refuse to participate in data relaying due to various reasons(such as resource limitations or social preferences).Meanwhile,misbehaving nodes may drop received messages but produce false reports to disrupt the data delivery process.Dealing with the non-cooperative nodes in M-SNs is extremely challenging because the misbehaving actions of selfish and malicious nodes are spread in space and time and the observation of one node might not sufficiently indicate the misbehavior of its encountered nodes.Several distributed algorithms and protocols have been developed to detect non-cooperative nodes and mitigate their impact on the data delivery per-formance.While the majority of existing methods primarily rely on the nodes' contact history,the role of nodes' individual and social preferences have not been explored sufficiently.Nev-ertheless,we believe that the nodes' social attributes and relationships(such as friendships and common interests)have long-term characteristics that can help predict their cooperation pref-erences swiftly,detect their possible selfish and malicious behaviors effectively,and stimulate their cooperation efficiently.In this dissertation,we develop efficient social-based protocols to combat selfish and mali-cious mobile nodes in non-cooperative MSNs.First,we provide an overview of non-cooperative MSNs in which we identify different selfish and malicious node behaviors and explore state-of-the-art solutions that aim to deal with non-cooperative nodes in MSNs.Next,we propose a set of utility-driven distributed mechanisms in which the nodes' individual and social routing util-ities are calculated and employed to detect their selfish behavior,distinguish their individually and socially selfish behavior,and promote their cooperation.We employed both analytical and simulation-based experiments to evaluate the performance of our proposed protocols where the nodes' mobility and social features are assigned using real-world datasets.In chapter 3,we propose a signalling game approach(Sig4UDD)to model interactions between well-behaved,individually selfish,and socially selfish nodes when the nodes prevent revealing their type and reactions to each other,which leads data forwarding under uncertain n-ode behavior.In Sig4UDD,we establish a belief system in which a node updates its belief about the type of its encountered nodes based on the properties of their forwarding messages that helps it predict their actual type(and thus their reactions)and make appropriate forwarding decisions.We apply Bayesian Nash equilibrium and perfect Bayesian equilibrium to respectively find the best strategies(i.e.,equilibrium points)in the one-stage and multi-stage node interactions and prove that a node would gain an optimal payoff given the strategy of another node.Our extensive experiments demonstrate that Sig4UDD outperforms some benchmark non-cooperative data for-warding protocols in terms of data delivery ratio and delay with minimum communication cost.In chapter 4,we propose a social-based watchdog scheme(SoWatch)in which watchdog nodes analyze messages received from their encountered nodes with respect to their social tie information in order to identify the nodes' selfish behavior in message relaying.Meanwhile,the watchdog nodes apply the second-hand watchdog information received from other nodes to improve the detection time and accuracy.Next,we design a reputation system in which watch-dog nodes identify selfish nodes based.on their direct and indirect watchdog information and distinguish individually and socially selfish nodes.Furthermore,we design a watchdog evalua-tion module to protect SoWatch against wrong watchdogs disseminated by malicious nodes in which a watchdog node investigates the truthfulness of the indirect watchdogs before applying them.Our experiments using real-world datasets illustrate that SoWatch outperforms a bench-mark contact-based watchdog system in terms of detection time by 45%and detection ratio by 10%with less communication overhead.In chapter 5,we propose a bargaining-based incentive scheme for social-aware routing pro-tocols(GISSO),which guarantees the faithfulness of selfish nodes in message relaying for other nodes and achieves a high end-to-end throughput.In GISSO,we distinguish high-beneficial and low-beneficial messages for each intermediate node based on their individual and social utili-ties.While an intermediate node is willing to relay the high-beneficial messages,we employ an alternating-offers bargaining game to analyze message trading between encountered nodes.In the bargaining process,the sender of a message negotiates with the receiver node over the value of its forwarding service in some rounds until they reach an agreement or the game finishes.We apply subgame perfect Nash equilibrium as the agreement of the players to maximize the utility of each player without decreasing the utility of another player.Reaching an agreement,the sender pays a certain amount of credit to the receiver to buy its forwarding service,and the receiver accepts to store and relay the message.The comparison of GISSO with some bench-mark routing protocols illustrates that GISSO dominates the selfish actions and outperforms the other algorithms in terms of data delivery ratio and delay with low communication overhead.