TY - JOUR
TI - Exploiting user interest similarity and social links for micro-blog forwarding in mobile opportunistic networks
AU - Allen, S.M.
AU - Chorley, M.J.
AU - Colombo, G.B.
AU - Jaho, E.
AU - Karaliopoulos, M.
AU - Stavrakakis, I.
AU - Whitaker, R.M.
JO - Pervasive and Mobile Computing
PY - 2014
VL - 11
TODO - null
SP - 106-131
PB - Elsevier B.V.
SN - 1574-1192
TODO - 10.1016/j.pmcj.2011.12.003
TODO - Smartphones;  Social aspects;  Social networking (online), Content dissemination;  Forwarding strategies;  Micro blogging;  Micro-blogging services;  Mobile;  Mobile opportunistic networks;  Pervasive;  Pervasive environments, Blogs
TODO - Micro-blogging services have recently been experiencing increasing success among Web users. Different to traditional online social applications, micro-blogs are lightweight, require small cognitive effort and help share real-time information about personal activities and interests. In this article, we explore scalable pushing protocols that are particularly suited for the delivery of this type of service in a mobile pervasive environment. Here, micro-blog updates are generated and carried by mobile (smart-phone type) devices and are exchanged through opportunistic encounters. We enhance primitive push mechanisms using social information concerning the interests of network nodes as well as the frequency of encounters with them. This information is collected and shared dynamically, as nodes initially encounter each other and exchange their preferences, and directs the forwarding of micro-blog updates across the network. Also incorporated is the spatiotemporal scope of the updates, which is only partially considered in current Internet services. We introduce several new protocol variants that differentiate the forwarding strategy towards interest-similar and frequently encountered nodes, as well as the amount of updates forwarded upon each encounter. In all cases, the proposed scheme outperforms the basic flooding dissemination mechanism in delivering high numbers of micro-blog updates to the nodes interested in them. Our extensive evaluation highlights how use can be made of different amounts of social information to trade performance with complexity and computational effort. However, hard performance bounds appear to be set by the level of coincidence between interest-similar node communities and meeting groups emerging due to the mobility patterns of the nodes. © 2011 Elsevier B.V.
ER -