Weight functions for shortest path routing of periodically scheduled burst flows

We investgate weighted shortest path routing mechanisms based on Dijkstra’s algorithm to find optimal paths for periodically scheduled burst flows in optical burst switching networks. Our objective is to reduce the flow blocking rate. We propose three dynamic weight functions to estimate the path level flow blocking probability based on wavelength utilization. We evaluate the performance of these weight functions using simulation. Results show that the weight function that incorporates both wavelength utilization and the number of inbound traffic sources has the best performance.