(Abridged) Based on XMM-Newton, Chandra and SDSS data, we investigate the baryon distribution in groups and clusters and its use as a cosmological constraint. For this, we considered a sample of 123 systems, with total masses in the mass range M500 = ~ 10^13 - 4 x 10^15 h_70^-1 Msun. The gas masses and total masses are derived from X-ray data under the assumption of hydrostatic equilibrium and spherical symmetry. The stellar masses are based on SDSS-DR8 data. For the 37 systems out of 123 that had both optical and X-ray data available, we investigated the gas, stellar and total baryon mass fractions inside r2500 and r500, and the differential gas mass fraction within the spherical annulus between r2500 and r500, as a function of total mass. For the other objects, we investigated the gas mass fraction only. We find that the gas mass fraction inside r2500 and r500 depends on the total mass. However, the differential gas mass fraction does not show any dependence on total mass for systems with M500 > 10^14 Msun. We find that the total baryonic content increases with cluster mass. This led us to investigate the contribution of the ICL to the total baryon budget for lower mass systems, but we find that it cannot account for the difference observed. The gas mass fraction dependence on total mass observed for groups and clusters could be due to the difficulty of low-mass systems to retain gas inside the inner region. Due to their shallower potential well, non-thermal processes are more effective in expelling the gas from their central regions outwards. Since the differential gas mass fraction is nearly constant it provides better constraints for cosmology. Using our total f_b estimates, our results imply 0.17 < Omega_m < 0.55.