This thesis uses a systems-level approach to study the cellular metabolism, unveiling new mechanisms and responses that were impossible to reach with traditional reductionists procedures. The results reported here have a potential application in areas like metabolic engineering and disease treatment. They could also be used in determining the accuracy of the gene essentiality of new genome-scale reconstructions. Different methods and techniques, within the contexts of Systems Biology and the field known as Complex Networks Analysis have been applied in this work to show different features of the robustness of metabolic networks. The specific issues addressed here range from pure topological aspec ts of the networks themselves to the balance of biochemical fluxes.