TWDM PON is a promising candidate to become the backhauling infrastructure of Fixed Mobile Converged (FMC) access-aggregation networks. One of the main features of a FMC network is that different types of traffic (that support, e.g., fixed and mobile network services) are transported over the same infrastructure, and the coexistence of different traffic types poses complex challenges in terms of capacity assignment as well as traffic management. If TWDM is the technology of choice, one question naturally raises: is it better to isolate different types of traffic over different wavelengths (i.e., Layer-1 convergence) or to share wavelength capacities among the different traffic types (i.e., Layer-2 convergence)? In the first case, to isolate the different types of traffic, we can assign logically-separated Virtual PONs (VPONs) to different traffic types, that will operate on disjoint sets of wavelengths. In the second case, different traffic types can share the same wavelengths in the network (i.e., a single convergent TWDM PON is used). We first model the problem resorting to a modified version of the classical bin packing problem. Then, through both numerical and qualitative considerations, we investigate the trade-offs between the two options under different scenarios in terms of traffic loads, user distribution, and network capacity.