End-to-end delay satisfaction balancing routing

This paper presents QoS-based routing algorithms using the end-to-end delay satisfaction balancing concept. The algorithms are suitable for networks that do not use per-path reservation. A nonlinear optimization problem and a gradient-based solution are formulated for off-line computation of the optimal route configuration. An approximation of the optimization problem is developed for on-line distributed processing. Using the approximation, vector routing tables can be used to set-up paths for arriving calls. Performance of the proposed schemes is evaluated and compared with that of minimum-delay, minimum-hop, and min-interference routing algorithms. Results show that using an objective function based on delay satisfaction balancing enables the network to accommodate more users of varying end-to-end delay requirements.