Grant: Discovery Grants Program - Individual

The telecommunication industry currently consumes 2-3% of global energy, and energy consumption constitutes 20-40% of the operating expenditure for mobile network operators. The energy footprint is especially high at the radio access network (RAN), where Base stations are a major source of energy consumption. The deployment of the 5G radio technologies is expected to further exacerbate the situation, hence raising major concerns on the environmental and economic sustainability of future-generation mobile infrastructures. The challenge must be tackled now, and primarily by developing RAN solutions that achieve high-energy efficiency and are sustainable from environmental and economic perspectives. In order to be credible, the aforementioned solutions must be designed in compliance with the specificities of virtualized RAN (vRAN) paradigms embraced by 5G, which combine RAN disaggregation, virtualization, and slicing support. In this context, this research program aims to explore the novel opportunities enabled by the zero-touch management of vRAN resources and functions towards more sustainable mobile communications. Specifically, my team and I will pursue three primary objectives: � Build open datasets of the power consumption at the User Equipment and in the vRAN infrastructure of a wide range of relevant Virtualized Network Functions (VNFs), and leverage these datasets to characterize the energy requirements of different VNF configuration scenarios. � Develop data-driven Network Intelligence (NI) algorithms based on machine-learning (ML) models for the anticipatory management and admission control of green network slices under diverse energy budgeting policies and in advance of fluctuations in the user demand. � Leverage optimization techniques and ML to develop new NI algorithms for the orchestration of vRAN resources over longer timescales for green slice VNF placement and chaining optimization as well as resource auto-scaling at both radio access and the cloud-based infrastructure. We will demonstrate the sustainability gains of the developed NI algorithms via emulations with large-scale measurement data as well as implementations and tests in a real 5G platform based on state-of-the-art Software Defined Radio tools. Designing a cleaner generation of telecommunication systems will contribute to reducing network operators' capital and operational expenditures, particularly at the radio access, eliminating network outages and optimizing network operations. The proposed program will produce the trainees required to drive the coming wave of future-generation virtualized mobile networks, an area where Canada has traditionally played a leadership role.