The 10-100 keV region of the electromagnetic spectrum contains the potential for a dramatic improvement in our understanding of a number of key problems in high energy astrophysics. A deep inspection of the universe in this band is on the other hand still lacking because of the demanding sensitivity (fraction of microCrab in the 20-40 keV for 1 Ms integration time) and imaging (~15" angular resolution) requirements. The mission ideas currently being proposed are based on long focal length, grazing incidence, multi-layer optics, coupled with focal plane detectors with few hundreds microns spatial resolution capability. The required large focal lengths, ranging between 8 and 50 m, can be realized by means of extendable optical benches (as foreseen e.g. for the HEXIT-SAT, NEXT and NuSTAR missions) or formation flight scenarios (e.g. Simbol-X and XEUS). While the final telescope design will require a detailed trade-off analysis between all the relevant parameters (focal length, plate scale value, angular resolution, field of view, detector size, and sensitivity degradation due to detector dead area and telescope vignetting), extreme attention must be dedicated to the background minimization. In this respect, key issues are represented by the passive baffling system, which in case of large focal lengths requires particular design assessments, and by the active/passive shielding geometries and materials. In this work, the result of a study of the expected background for a hard X-ray telescope is presented, and its implication on the required sensitivity, together with the possible implementation design concepts for active and passive shielding in the framework of future satellite missions, are discussed.