Load balancing in multi-beam satellite systems
Gun Akkor, John S. Baras and Michael Hadjitheodosiou
Number: CSHCN TR 2004-9, Year: 2004, Advisor: John S. Baras
We propose an approach to optimize resource sharing and in a multi-beam broadband satellite system that supports both unicast and multicast flows. We show that in this architecture, the load on every spot-beam queue could be different, depending on the type of the flows and the distribution of the receivers across spot-beam coverage areas. This load unbalance may significantly under-utilize the system resources and decrease the system throughput when both unicast and multicast flows are active in the system. In this paper, we formulate an optimization problem for intra- and inter-beam resource sharing such that the variance of the session rates experienced by users of a flow located in different beam coverage areas is minimized. The result of our resource allocation also determines the maximum sustainable rate of each flow. We calculate the beam utilization and maximum sustainable sessions rates with and without optimization and compare the results. We conclude that this method significantly improves the session rates and overall utilization of the system when both unicast and multicast flows are active.