Existing information risk management methodologies, as a rule, come down to determining the optimal set of countermeasures that maximize risk reduction under specified constraints. At the same time, it is extremely rare to take into account the possible cumulative effect of the use of countermeasure complexes and almost never the mutual influence of risks on each other, which takes place in any complex system. The proposed paper discusses one of the possible types of such influence caused by the structure of the information system. The main question is to what extent the latter will affect the overall risk of the system. The result is an a priori estimate of the possible variation in the magnitude of the integral risk of the system with different placements of its elements in a given tree structure with a single-vertex perimeter. In subsequent studies, this result is supposed to be generalized to the case of an arbitrary topology structure.