Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A Survey on AI-driven Energy Optimisation in Terrestrial Next Generation Radio Access Networks

Kishan Sthankiya, Nagham Saeed, Greg McSorley, Mona Jaber, Richard G. Clegg

TL;DR

It is found that extracting even rough estimates of the operational energy cost of AI models and data processing pipelines is complex, which hinders a meaningful comparison between the energy savings from AI techniques and their associated energy costs.

Abstract

This survey uncovers the tension between AI techniques designed for energy saving in mobile networks and the energy demands those same techniques create. We compare modeling approaches that estimate power usage cost of current commercial terrestrial next-generation radio access network deployments. We then categorize emerging methods for reducing power usage by domain: time, frequency, power, and spatial. Next, we conduct a timely review of studies that attempt to estimate the power usage of the AI techniques themselves. We identify several gaps in the literature. Notably, real-world data for the power consumption is difficult to source due to commercial sensitivity. Comparing methods to reduce energy consumption is beyond challenging because of the diversity of system models and metrics. Crucially, the energy cost of AI techniques is often overlooked, though some studies provide estimates of algorithmic complexity or run-time. We find that extracting even rough estimates of the operational energy cost of AI models and data processing pipelines is complex. Overall, we find the current literature hinders a meaningful comparison between the energy savings from AI techniques and their associated energy costs. Finally, we discuss future research opportunities to uncover the utility of AI for energy saving.

A Survey on AI-driven Energy Optimisation in Terrestrial Next Generation Radio Access Networks

TL;DR

It is found that extracting even rough estimates of the operational energy cost of AI models and data processing pipelines is complex, which hinders a meaningful comparison between the energy savings from AI techniques and their associated energy costs.

Abstract

This survey uncovers the tension between AI techniques designed for energy saving in mobile networks and the energy demands those same techniques create. We compare modeling approaches that estimate power usage cost of current commercial terrestrial next-generation radio access network deployments. We then categorize emerging methods for reducing power usage by domain: time, frequency, power, and spatial. Next, we conduct a timely review of studies that attempt to estimate the power usage of the AI techniques themselves. We identify several gaps in the literature. Notably, real-world data for the power consumption is difficult to source due to commercial sensitivity. Comparing methods to reduce energy consumption is beyond challenging because of the diversity of system models and metrics. Crucially, the energy cost of AI techniques is often overlooked, though some studies provide estimates of algorithmic complexity or run-time. We find that extracting even rough estimates of the operational energy cost of AI models and data processing pipelines is complex. Overall, we find the current literature hinders a meaningful comparison between the energy savings from AI techniques and their associated energy costs. Finally, we discuss future research opportunities to uncover the utility of AI for energy saving.

Paper Structure

This paper contains 19 sections, 7 equations, 4 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Related Survey Papers
  3. Scope and Contributions
  4. NG-RAN Architecture
  5. RAN Power Consumption Modeling
  6. Analytical Models
  7. Empirical Models
  8. Power modeling summary
  9. Energy Saving Techniques
  10. Time-domain
  11. Frequency-domain
  12. Power-domain
  13. Spatial-domain
  14. Summary of AI techniques for RAN efficiency
  15. Operational Energy Cost of AI and ML
  16. ...and 4 more sections

Figures (4)

  • Figure 1: High-level taxonomy of topics covered in this survey
  • Figure 2: Overview of 5G System. Composed of UE, NG-RAN (shaded) and 5G Core Network.
  • Figure 3: A comparison of power consumption models from the literature focused on the radio unit, base station and Massive Multiple-Input and Multiple-Output ( mMIMO ) system. Components include the baseband-digital front end (BB/DFE), baseband unit (BBU), channel coding and decoding (COD/DEC), mains power supply losses (MS), direct current conversion losses (DC-DC), active cooling losses (COOL), radio frequency transceiver (RF TRX), power amplifier (PA) and user equipment (UE). Grey components without a dot indicate a component with load-independent power consumption. Components with a dot represent a dynamic power consumption, where the color represents the influencing factor.
  • Figure 4: Taxonomy of power saving techniques in this survey