The article discusses the current task of optimal resource allocation in CLOUD systems that support hybrid cloud data storage. The purpose of this article is to develop an optimal distribution method of several heterogeneous transactions to a hybrid cloud data storage, connected by a common bandwidth limitation. The optimization criterion will be the minimum cost. Information flows have certain points of departure and destination, are heterogeneous flows with common restrictions on the bandwidth of the communication channels used. The results obtained. The method is based on the construction of network graphs. The optimization problem is reduced to the distribution of flows in the network in such a way that, if the requirements of the cloud hybrid data storage are met, the cost of transmitting the flow in the network is minimal. To solve it, an iterative algorithm for constructing the maximum admissible flow is proposed. At each iteration, the simplex table of network graphs is modified. The direction of further research is the development of a method for optimizing the structure of cloud data storage.