Solid-state NMR is emerging as a method for resolving structural information for large biomolecular complexes, such as membrane-embedded proteins. In principle, there is no molecular weight limit to the use of the approach, although the complexity and volume of data is still outside complete assignment and structural determinations for any large (Mr > approx 30,000) complex unless specific methods to reduce the information content to a manageable amount are employed. Such methods include specific residue-type labeling, labeling of putative segments of a protein, or examination of complexes made up of smaller, manageable units, such as oligomeric ion channels. Labeling possibilities are usually limited to recombinant or synthesized proteins, and labeling strategies often follow models from a bioinformatics approach. In all cases, and in common with most membrane studies, sample preparation is vital, and this activity alone can take considerable effort before NMR can be applied--peptide or protein production (synthesis or expression) followed by reconstitution into bilayers and resolution of suitable sample geometry is still technically challenging. As experience is gained in the field, this development time should decrease. Here, the practical aspects of the use of solid-state NMR for membrane protein structural determinations are presented, as well as how the methodology can be applied. Some successes to date are discussed, with an indication of how the area might develop.