In the current paper a question of the resource-saving tasks distribution is under consideration. The problem of computational resource saving is topical because of the enormous data volumes, which are preprocessed partially by the fog- and edge- network layers. In general, scheduling and resource allocation are modeled via combinatorial optimization problems without consideration of the fact that the computational environment is geographically distributed. The consequence of such distribution is that the tasks assigned to some nodes have to transmit the data through some transit network sections. As the data transmission produces workload and consumes time, which degrade the average residual time of the nodes, in this paper we propose the novel problem model, which is structural-parametric and focuses not only on the functional tasks assignment to the nodes, but to the data transmission workload, which disseminates through the data transmission routes. The generic solution method is proposed on the base of multiplicative convolution and random search. The produced results show the positive effect of the workload distribution on the nodes reliability function values.