On the Bounded-Delay Minimum-Cost Path Problem of Quality-of-Service Routing

A wide range of communication-intensive real-time multimedia applications over networks require guaranteed end- to-end quality of service (QoS). QoS-routing in networks aims to find feasible routes that have sufficient resources to satisfy imposed QoS-constraints. The bounded-delay minimum-cost path problem of QoS-routing is provably difficult; several efficient heuristic algorithms exist for computing feasible solutions to the problem. We design and implement an improved heuristic algorithm for solving the problem, which employs bounded-size buffers for properly maintaining competitive cost-delay combinations for feasible-path extension and offers a better chance to find optimal/quality solutions than the known heuristics. We also prove its correctness of computing feasible solutions in worst-case polynomial-time and -space complexities. Two sets of experimental results via simulations are presented: a tradeoff between the cost/time-performance of the algorithm and buffer size, and a delay/cost/time-performance comparison study.