It is now generally acknowledged that microbial populations will be present within nuclear waste repositories and that the consequences of such activity on repository performance must be assessed. Various modelling approaches--based either on mass balance/thermodynamics or on kinetics--have been developed to provide scoping estimates of the possible development of these populations. Past work has focused on particular areas of the repository which can be considered relatively homogeneous and hence can be represented by some kind of 'box' or 'mixing tank'. In reality, however, waste repositories include a range of engineering materials (steel, concrete, etc.) which are emplaced at depth in a rock formation. Strong chemical gradients--of the type which may be exploited by lithoautotrophic microbial populations--are likely to be found at the contacts between different materials and at the interface between the engineered structures and the host rock. Over the long timescales considered, solute transport processes will cause the locations of strong chemical gradients to move, forming reaction fronts. The high-pH plume resulting from the leaching of cement/concrete in some repository types is a particularly important example of such a reaction front. Redox fronts, which may occur in different areas of all kinds of repositories, also play an important role and would be locations where microbial activity is likely to be significant. In this paper, the key microbial processes expected at (or around) interfaces and fronts will be discussed, with particular emphasis on the development of quantitative models. The applicability of the models used wil be tested by considering similar fronts which can be found in natural systems.