Online social networks have gained importance in recent years. Furthermore, there is a need for designing smart applications for social networks which take into account the behaviour of dynamical processes over these networks. This requires structural knowledge of the network helpful in understanding the dynamical process. Here we study a broad category of such processes called rumor spreading processes. We simulate a typical rumor spreading scenario on a real social network graph of 5.2 million nodes and 72 million edges. We compare the results of this simulation to two synthetically generated Erd ̋s-R ́nyi [1] and o e power-law random graphs. Our simulation shows that the behavior of the rumor spreading process is considerably different in social networks as compared to the one observed on above mentioned synthetic random graphs. These simulations have possible implications to applications in viral advertising, social marketing, worm attacks, online political campaigns, peer-to-peer communication networks.