Routing Protocols And Incentive Mechanisms In Socially Aware Networking

Socially aware networking (SAN) is an emerging paradigm in intermittently connected networks that attracted significant attention by researches recently. In this paradigm, there is a lack of end-to-end connections between mobile devices and they communicate with each other via wireless transmission technologies such as Bluetooth or WiFi opportunistically. Generally, mobile devices are carried by their owners, hence, they mirror mobility patterns, social characteristics and behaviors of the users. Consequently, the social properties of the mobile devices and users can be exploited to design efficient networking protocols in pervasive mobile environments.Data forwarding in SAN paradigm is a challenging problem since mobile nodes use store-carry-and-forward scheme to deliver messages between interested users. Therefore, the efficiency of data delivery heavily relies on the contact prediction as well as the cooperation level of the mobile nodes. In order to tackle these two issues, various social network analysis concepts such as social tie, community structure, social similarity, and mobility patters have been utilized to choose the best intermediate nodes. In addition, several incentive schemes have been devised in order to stimulate network nodes to cooperate in data relaying for other network nodes.In this dissertation, we study state-of-the-art data routing and forwarding protocols in SAN paradigm comprehensively and present four major contributions. We propose two social-aware routing protocols namely BEEINFO and PIS in which social properties of mobile users are used to improve data forwarding performance in SAN paradigm. Furthermore, CAIS and Com-BIS schemes are designed to stimulate selfish mobile nodes to cooperate for other network nodes.In summary, the major contributions of this dissertation can be summarized as follows:1. Community structure is one of the popular concepts to analyze social relationships between mobile nodes in SAN. However, the topology of mobile network in unstable and changes dynamically, which makes it difficult to construct and maintain. We present a set of routing protocols called BEEINFO to perceive the dynamic environments whenever necessary. In BEEINFO, nodes with similar interests on data are grouped into the same community. Nodes imitate the behavior of bees and perceive the density information of passed communities and social tie information of nodes in same community when they are moving. By use of these information, BEEINFO make the forwarding decision. Simulation results show that BEEINFO obtain higher routing efficiency compared to Epidemic and PROPHET.2. Various social characteristics and relationships can be available between the nodes which can affect data forwarding in different ways. In order to obtain more stable forwarding efficiency, we present a multidimensional routing protocol called PIS. It combines three social factors including physical proximity, data interest and social relationship to make a comprehensive forwarding decision. Additionally, PIS considers time factor and maintain a time slot management to discover users’movement similarities in different time periods in a day. Simulations are based on datasets of SIGCOMM2009and INFOCOM2006, and results show that PIS performs more stable and higher forwarding performance.3. The existence of selfish nodes degrade the performance of data forwarding in SAN paradigm significantly. We propose a Copy Adjustable Incentive Scheme (CAIS) to stimulate selfish nodes to cooperate and relay data for others. CAIS combines cooperative degree of nodes with nodes’requirements for higher delivery efficiency. Nodes with higher cooperative degree can replicate their favorite messages through the network more than others which increase their routing performance. CAIS adopts social and non-social credits to manage the cooperation behaviors between the nodes inside and outsider a community. Additionally, CAIS utilizes single source distribution mode to simplify the credit distribution management. Simulation results illustrate that CAIS stimulate selfish nodes effectively and obtain higher delivery ratio, lower overhead and shorter average latency.4. Barter-based incentive schemes have been extensively used by researchers to address fairness issues in SAN paradigm. However, data exchange process using a barter-based method is challenging when two encounter nodes do not have the same number of messages to exchange. We propose a tit-for-tat (TFT) based incentive mechanism named Com-BIS which barters forwarding services between nodes in different communities. Data forwarding problem in Com-BIS is considered as a forwarding resource allocation problem. Simulation experiments are conducted on the dataset of INFOCOM2006and the results show that, Com-BIS not only stimulates selfishness nodes to cooperate successfully, but also improves the forwarding efficiency effectively.