RUDN University Mathematicians Propose a Platform for 6G Networks

MOSCOW, Russian Federation - Sept. 17, 2021 - PRLog -- RUDN University mathematicians together with colleagues from China, Egypt, Morocco, and Russia have proposed a platform for optimizing the loads in mobile networks, which can be used in sixth-generation networks (6G). The first results showed that the platform can increase the efficiency of the equipment by 30%. The results are published in the journal Simulation Modeling Practice and Theory.

The fifth-generation (5G) mobile networks have only recently launched in several countries, but it is already clear that they will not match with the growing demands. The Research Center of 5G Wireless Networks Simulation at the RUDN University has already started advanced development of the next generation of mobile networks (6G). One of the disadvantages of 5G is the "classic", static network topology. This means that the configuration of the equipment (network nodes) does not change, and the load on different network nodes is not redistribute. RUDN University mathematicians have proposed a fundamentally new approach — a dynamic network topology.

"With the development of networks and a smooth transition to 5G & 6G networks, the number of subscribers' traffic will grow exponentially. Standard methods for improving communication channels are no longer working. Therefore, it is necessary to vector development of networks towards dynamically configurable systems", said Abdukodir Khakimov, the junior researcher at the Research Center of 5G Wireless Networks Simulation, Institute of Applied Mathematics & Communication Technology, RUDN University.

The system proposed (https://www.sciencedirect.com/science/article/pii/S156919...) by mathematicians got called REx (Resource Exchange). Its essence lies in the fact that processes are redistributed between network nodes depending on the load at the moment. The network architecture consists of three elements — cloud computing, the network operator, and the REx platform itself. The platform can operate in two modes — automatic and on-demand placement. In the automatic mode, the first step is the traffic analysis. It provides recommendations for optimal application placement, then the necessary capacities are allocated for them. In the second mode, traffic analysis does not occur automatically, but by the network operator. The platform also can predict future traffic to optimize the distribution of capacity across nodes in advance.

RUDN University mathematicians conducted a number of simulations to check the REx platform, and also started to test the platform together with one of the telecom operators and several third-party services. In a real experiment, during peak periods, the load dropped to 30%. In the future, scientists plan to improve the methodology of traffic forecasting. To do this, mathematicians want to introduce machine learning methods and the ARIMA (Autoregressive Integrated Moving Average Model) time series analysis method into REx.