Performance study of adaptive routing algorithms for LEO satellite constellations under Self-Similar and Poisson traffic

A comparative study of routing techniques is carried out for LEO constellations interconnecting high speed terrestrial networks assuming Poison and Self-Similar input traffic. Shortest path routing as well as optimal routing (flow deviation) methods are applied for balanced and unbalanced traffic load and for uniform and non uniform distribution of the earth stations. The performance of flow deviation method is proved to be very successful even for the LEO complicated networks for both Poisson or Self-Similar input. A modification of the classic flow deviation algorithm regarding the number of paths we work with is proposed. Indeed, it is proved that a k-paths flow deviation method is always easy to obtain and gives robust results for any traffic pattern at a very affordable algorithmic complexity.