Dipartimento di Ingegneria informatica, automatica e gestionale

In this paper, a distributed, non-cooperative and dynamic load-balancing algorithm is proposed in the context of multi-commodity adversarial network equilibria with constrained providers’ capacities. The algorithm is proven to converge to a generalised Wardrop user-equilibrium, referred to as Beckmann equilibrium in the literature, in which, for each commodity, the latencies of the unsaturated providers are equalized. The algorithm is then used as a Multi-connectivity algorithm in the context of 5G heterogeneous networks, in which the user equipments are able to use different access networks simultaneously to increase the transmission capacity and/or to improve the transmission reliability. The proposed controller provides a solution for dynamic traffic steering by distributing the traffic load over the available heterogeneous access points, considered as capacity providers. Simulation results validate the approach. The developed network simulator is available as an open-source environment De Santis et al. (2020).