This paper researches the building principles and administration of virtual data processing centers based on hyper-converged systems over OpenFlow. We provide the implementation features of such virtual centers on the basis of software-defined networking that is managed by a dedicated controller (a server). We suggest the graph administration model of hyper-converged system resources compliant with required performance on the one hand and economic requirements on the other. Based on the proposed model, the implementation of a greedy control algorithm for the virtual data processing center over OpenFlow was examined. This algorithm assigns the requests to physical resources by using of dedicated server software. The advantages of such hyper-converged system model on performance issues were outlined, e.g., multi-threaded routing and security, elimination of the majority of current threats. We summarize the possibilities of transition to network infrastructure in these virtual data processing centers. Such infrastructure is focused on data and usage of blockchain technology providing high reliability and content protection.