Scalable Micro-blogging in Mobile Social Communities

Micro-blogging in Mobile Social Network can be very successful because of spatiotemporal properties of user interests and generated messages. Users can follow their interests and receive messages of their interests without getting overwhelmed instead of having to identify and follow users having similar interests. We propose a distributed and scalable micro-blogging protocol which exploits community structure and heterogeneous popularity of nodes. Control messages across communities are restricted by aggregating control information and by limiting storage of state information at hub/gateway nodes. The protocol dynamically estimates distance up to which interest in the message is high and restricts message forwarding accordingly. Simulation results show that our protocol performs far better than existing protocol. We also propose three novel models to generate user interest profiles synthetically.