RESOURCE DIMENSIONING OF BROADBAND SATELLITE RETURN NETWORKS AFFECTED BY RAIN FADE

The use of Ka-band in satellite links has made rain attenuation a major concern in satellite network design. Fade mitigation techniques come at the expense of higher satellite resource consumption, such as bandwidth and power. An accurate estimation of this consumption is essential for the satellite service providers' business case, product strategy, and overall service pricing. However, the spatial correlation of rain fade introduces a high level of model complexity, and no method is currently available to compute its impact on resource consumption. Focusing on the return link of satellite broadband networks, this dissertation proposes a satellite resource dimensioning process that accounts for such a correlation in several scenarios, depending on the network's adaptability to rain fade.

Firstly, we investigate nonadaptive network scenarios and answer the following question: how can the long-term bandwidth requirement of the network be minimized, given a set of ground terminals, modulations and codings, and discrete bandwidths? We formally define the long-term carrier allocation problem and analyze current practical solutions. We subsequently investigate two other potential solutions, found to be more bandwidth-efficient: one based on heuristics and another based on mixed integer linear programming. Finally, we look at the impact of several parameters on the performance of those three methods. Overall, we observe marginal reductions in bandwidth, however, significant (>10%) gains are reached for networks with small return links with low committed information rates.

Secondly, we investigate semi-adaptive network scenarios with the introduction of adaptive coding and modulation. However, these technologies come at the cost of higher complexity when designing the network's carrier plan and user terminals. Taking into account those issues is even more important when the satellite link uses frequencies in Ka-band and above, where rain attenuation is a major concern. To consider such phenomena, we reformulate the previously presented solutions to factor in spatially correlated attenuation time series, in the form of a mixed integer linear programming optimization problem. The numerical results for a test scenario in Europe show significant bandwidth improvements.

Lastly, we investigate fully-adaptive network scenarios and introduce multibeam aspects. We formulate a quantile estimation problem based on the broadband service level agreements. Then, we solve this problem for a given confidence relative interval using spatially correlated rain fade sample generators. Finally, we provide numerical results for residential and enterprise broadband satellite scenarios, allowing us to determine the underestimation and overestimation of satellite resource consumption made by optimistic (independent) and pessimistic (fully correlated) rain fade assumptions, respectively. Results show that for both assumptions, the satellite resource consumption can be significantly underestimated or overestimated, thus proving the importance of considering the spatial correlation of rain fade in the satellite resource dimensioning problem.