Social-aware data dissemination service in mobile social network with controlled overhead

The popularity of Mobile Social Networks (MSNs) paves a way for mobile users to download or view their interested contents not only from access points of Internet but also from other mobile users through short-range radio communications like Bluetooth and Wi-Fi. Through communications between users in MSNs, the bandwidth of short-range communications can be better utilized and relieve the stress on the bandwidth of cellular network. However, due to intermittent connectivity in MSNs, nodes in MSNs can only forward messages opportunistically. Therefore, how to disseminate data accurately and efficiently turns out as a challenging problem. Most previous works focus more on increasing delivery ratio while paying less attention on reducing overhead measures the number of copies that are stored and spread by relay nodes that are not interested at all in the contents. Note that high data dissemination ratio is usually accompanied by high and uncontrolled overhead, resulting in heavy storage burden on the nodes in MSNs and low quality of data dissemination. This motivates us to propose a new data dissemination scheme to maximize data dissemination ratio with controlled overhead. And more specifically, with given limitation on the overhead of spreading the messages, we design a new data dissemination scheme to maximize data dissemination ratio. In our data dissemination scheme, a time-homogeneous Markov model is designed to analyze the interest transitions of every node's neighbors to decide when to forward the message and which node is chosen to forward the message further. Furthermore, two utility functions are investigated to evaluate the service ability of nodes for each kind of interest among the messages. The experimental results on simulated and real datasets demonstrate that our new scheme can outperform existing protocols with higher delivery ratio and lower delivery latency under controlled overheads.