Bandwidth resource allocation in integrated access and backhaul networks

Abstract

One of the promising technologies that allows currently deployed 5G and the anticipated 6G networks to cope with the ever increasing demand for high throughput low latency data services is Integrated Access and Backhaul. Self driving cars, augmented reality games and large scale data streaming are simple examples of new applications that require a large amount of low latency traffic. Integrated Access and Backhaul can provide better service to such applications by extending the traditional cellular access and combining access and backhaul resources. However, the actual performance gain depends on the specific allocation of the radio resources. In this paper we address this challenge and study new ways to allocate bandwidth across such networks. We formulate the IAB Resource Allocation Problem (IABrap) and provide a novel approximation algorithm with guaranteed performance to solve it. We also study a new, ML method, that is based on applying GNN (Graph Neural Network) to this problem. We evaluate the expected performance of both methods in realistic scenarios using a self-developed network simulator. Our results indicate that combining traditional algorithmic techniques with state of art ML can provide better practical algorithms.