Tomography has been applied to holographic imaging systems recently to improve the 3D imaging performance. However, there are two distinct ways to achieve this: either by rotation of the object or by rotation of the illumination beam. We provide a transfer function analysis to distinguish between these two techniques and to predict the 3D imaging performance in holographic tomography when diffraction effects are considered. The results show that the configuration of rotating the illumination beam in one direction while fixing the sample leads to different 3D imaging performance as compared to the configuration of rotating the sample. The spatial frequency cutoff is nonisotropic in the case of rotating the illumination, and a curved line of singularity is observed. Rotating of the sample, on the contrary, has more symmetry in spatial frequency coverage but has a single point of singularity. The 3D transfer function derived can be used for 3D image reconstruction.