The evolution of 5G networks towards beyond 5G (B5G) and 6G, is expected to support emerging use-cases such as holographic telepresence and extended reality, which may require terabit-per-second throughput, sub-millisecond latency, and/or ultra-high reliability. Network slicing has been envisaged as a key enabler to satisfy these diverse requirements, by creating multiple isolated end-to-end virtual networks dedicated to different services, on top of a common physical infrastructure. To support the multi-faceted requirements of emerging applications, B5G and 6G warrant a highly flexible end-to-end network architecture, enabled by network slices spanning both the mobile core as well as the radio access network (RAN). However, legacy RAN solutions rely on closed-box functions and proprietary interfaces, making it difficult to slice the RAN, thus impeding the vision of end-to-end network slicing.
The Open RAN paradigm aims at alleviating some of the difficulties by leveraging RAN disaggregation, virtualization, radio network intelligence, and open interfaces. Adoption of the Open RAN paradigm is expected to yield lower operational expenditures due to the energy-efficient placement of the RAN functions and statistical multiplexing gains as a result of centralized baseband processing. However, Open RAN is not free of challenges; orchestrating and managing disaggregated RAN functions using a common orchestration platform and steering traffic through the fronthaul, midhaul and backhaul network segments are just a few of the research challenges impeding widespread adoption of Open RAN principles. In this project, we will leverage lessons learned from cloud-native systems as well as dynamic radio resource allocation algorithms powered by artificial intelligence (AI) to deploy a testbed based on Open RAN principles that will allow for the realization of end-to-end network slicing.
M. Mushtaq, M. Golkarifard, N.Shahriar, R. Boutaba, A. Saleh. Optimal Functional Splitting, Placement and Routing for Isolation-Aware Network Slicing in NG-RAN. IEEE/ACM/IFIP Conference on Network and Service Management (CNSM). Niagara Falls, Canada, October 30 - November 2, 2023.
M. Zangooei, M. Golkarifard, M. Rouili, N. Saha, and R. Boutaba. Flexible RAN Slicing in Open RAN with Constrained Multi-agent Reinforcement Learning. IEEE Journal on Selected Areas in Communications - Special issue on Open RAN. Accepted September, 2023.
M. Zangooei, N. Saha, M. Golkarifard and R. Boutaba. Reinforcement Learning for Radio Resource Management in RAN Slicing: A Survey. IEEE Communications Magazine. IEEE Press. Vol. 61(2), pp. 118-124, February 2023.