How well does your encounter-based application disseminate information?

In real-world, humans exhibit heterogeneous mobility patterns and are typically interested to receive context-based localized information originating from different data sources. Current works focus on sequential single pair-wise contacts among people under homogeneous mobility and single data source. This restricted evaluation for only single-scenarios do not collectively consider real-world mobility aspects of data dissemination. This approach based on narrow mobility aspects is mainly imposed due to intrinsic complexity of real-world. Deploying real experiments or simulating these real-world scenarios can be exhaustive and impractical. To include real-world mobility aspects and allow the understanding of scenarios bounds, we present a Markov model that collectively considers them to predict the performance of data dissemination. Our model allows multiple simultaneous pair-wise contacts among people under heterogeneous mobility and predicts the upper bound of data dissemination time. We achieve tighter upper bound of dissemination time than existing approaches by utilizing the long tail cut-off property of inter contact time distribution and data gathering process. We validate our model through different real-world traces from diverse environments and obtained the upper bound of data dissemination time with 5-10% error. We believe our model allows the pre-deployment performance analysis of encounter-based applications.