Examines the system's ability to meet real-time communication constraints with a degraded satellite constellation.

This thesis provides a performance evaluation of the IRIDIUM Low Earth Orbit Satellite system. It examine's the system's ability to meet real time communications constraints with a degraded satellite constellation. The analysis is conducted via computer simulation. The simulation is run at low, medium, and high loading levels with both uniform and non-uniform traffic distributions. An algorithmic approach is used to select critical satellites to remove from the constellation. Each combination of loading level and traffic distribution is analyzed with zero, three, five and seven non-operational satellites. The measured outputs are ene-to-end packet delay and packet rejection rate. In addition to the delay analysis, a user's ability to access the network with a degraded satellite constellation is evaluated. The average number of visible satellites, cumulative outage time, and maximum continuous outage time are analyzed for both an Equatorial city and a North American city. The results demonstrate that the IRIDIUM network is capable of meeting real-time communication requirements with several non-operational satellites. Both the high loading level and the non-uniform traffic distribution have a significant effect on the network's performance. The analysis of both network delay performance and network access provides a good measure of the overall network performance with a degraded satellite constellation.

This research presents a first of its kind comparative analysis of the Extended Bellman-Ford and Darting algorithms, using the Iridium low earth orbit (LEO) satellite system configuration for the simulation environment. The algorithms are compared to one another via discrete-event computer simulation and evaluated based on their ability to route real-time voice communications under low, medium, and high network loading conditions. The algorithms' ability to meet real-time voice constraints is evaluated with a full and degraded satellite constellation using an algorithmic satellite removal method. The investigation results indicate that both algorithms are suitable for use in a LEO environment and are capable of meeting the real-time voice communications requirements as long as a load-balancing mechanism is in place to route traffic around heavily loaded satellites. The results also indicate that the Iridium system is robust, capable of meeting the real-time voice constraints even when the constellation is degraded.