Game-theoretic model of asymmetrical multipath selection in pervasive computing environment

In order to transfer the ever increasing multimedia data we need to make use of multiple paths to realize the bandwidth aggregating. In pervasive computing environment, the combination of ubiquitous overlay networks and the capacity of multihoming can provide the possibility of exploiting plentiful available paths. Thus, we introduce a Pervasive Multipath Architecture (PEMA), in which we use the game theory to investigate the selfish strategic collaboration of multiple heterogeneous overlays when they are allowed to use the massively-multipath transfer. Overlay networks are modeled as “players” in this multipath selection game, and we study the asymmetric case where all overlays have different Round Trip Times (RTT) and different degrees of waste. We demonstrate the existence and uniqueness of Nash equilibrium (NE). In addition, we find that overlays differing only in their RTTs still receive proportional throughput shares and utilities at the NE. However, if overlays differ only in their degrees of waste, a more wasteful overlay has a larger utility and a larger throughput (bandwidth) share than a less wasteful overlay. Since maintaining connected paths consumes resource, we further consider a more generalized cost function where an overlay's total resource consumption includes both transferring data packets and maintaining path connections. In this general game, we find that the overlay is more conservative, which opens smaller number of paths and obtains smaller efficiency loss at the NE than in other simplified games. Our simulations confirm the effectiveness and friendliness of multipath transfer for a range of path numbers and in the presence of multi-overlay traffic.