Dynamic capacity allocation for quality-of-service support in IP-based satellite networks

This thesis concentrates on advanced dynamic capacity allocation (DCA) schemes that allow an efficient and fair sharing of the BSA return channel capacity among many user terminals (UTs), by dynamically adapting each UT's capacity assignment to the input traffic variations, while satisfying QoS requirements. DCA performance is evaluated with respect to satellite bandwidth utility (utilization) and QoS performance experienced by user traffic (expressed in terms of packet delay and loss). DCA performance is affected by various elements/parameters, some internal to the DCA scheme and others, external. Relying on simulation, the thesis considers the effects of these elements/parameters on performance, focusing on their optimization in order to balance the DCA performance against the signaling overhead and computational/algorithmic complexity of the DCA scheme. A novel DCA scheme is proposed, capable of providing Differentiated Services (DiffServ) IP QoS support for multimedia applications and high bandwidth utility, while maintaining reduced signaling overhead and computational/algorithmic complexity. (Abstract shortened by UMI.