Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

xPUE: Extending Power Usage Effectiveness Metrics for Cloud Infrastructures

Guillaume Fieni, Romain Rouvoy, Lionel Seinturier

TL;DR

This paper defines xPUE, a family of end-to-end energy-efficiency metrics that extend the traditional PUE to include hardware and software layers in cloud infrastructures. It formalizes hpue and vpue to quantify energy per unit of computation across hardware and software, and introduces cpue as a multiplicative combination across nested cloud layers. Implemented via the PowerAPI toolkit and SmartWatts, xPUE enables real-time, per-process energy disaggregation and supports deployment in containerized environments. Empirical validation across OpenStack and Kubernetes platforms shows that idle states, cooling choices, hardware configurations, and control-plane workloads significantly affect end-to-end efficiency and carbon implications, underscoring the need for holistic optimization of both hardware and software layers. The work concludes with concrete recommendations and demonstrates the potential of xPUE to guide energy-aware design and operation of cloud services at scale.

Abstract

The energy consumption analysis and optimization of data centers have been an increasingly popular topic over the past few years. It is widely recognized that several effective metrics exist to capture the efficiency of hardware and/or software hosted in these infrastructures. Unfortunately, choosing the corresponding metrics for specific infrastructure and assessing its efficiency over time is still considered an open problem. For this purpose, energy efficiency metrics, such as the Power Usage Effectiveness (PUE), assess the efficiency of the computing equipment of the infrastructure. However, this metric stops at the power supply of hosted servers and fails to offer a finer granularity to bring a deeper insight into the Power Usage Effectiveness of hardware and software running in cloud infrastructure.Therefore, we propose to leverage complementary PUE metrics, coined xPUE, to compute the energy efficiency of the computing continuum from hardware components, up to the running software layers. Our contribution aims to deliver realtime energy efficiency metrics from different perspectives for cloud infrastructure, hence helping cloud ecosystems-from cloud providers to their customers-to experiment and optimize the energy usage of cloud infrastructures at large.

xPUE: Extending Power Usage Effectiveness Metrics for Cloud Infrastructures

TL;DR

This paper defines xPUE, a family of end-to-end energy-efficiency metrics that extend the traditional PUE to include hardware and software layers in cloud infrastructures. It formalizes hpue and vpue to quantify energy per unit of computation across hardware and software, and introduces cpue as a multiplicative combination across nested cloud layers. Implemented via the PowerAPI toolkit and SmartWatts, xPUE enables real-time, per-process energy disaggregation and supports deployment in containerized environments. Empirical validation across OpenStack and Kubernetes platforms shows that idle states, cooling choices, hardware configurations, and control-plane workloads significantly affect end-to-end efficiency and carbon implications, underscoring the need for holistic optimization of both hardware and software layers. The work concludes with concrete recommendations and demonstrates the potential of xPUE to guide energy-aware design and operation of cloud services at scale.

Abstract

The energy consumption analysis and optimization of data centers have been an increasingly popular topic over the past few years. It is widely recognized that several effective metrics exist to capture the efficiency of hardware and/or software hosted in these infrastructures. Unfortunately, choosing the corresponding metrics for specific infrastructure and assessing its efficiency over time is still considered an open problem. For this purpose, energy efficiency metrics, such as the Power Usage Effectiveness (PUE), assess the efficiency of the computing equipment of the infrastructure. However, this metric stops at the power supply of hosted servers and fails to offer a finer granularity to bring a deeper insight into the Power Usage Effectiveness of hardware and software running in cloud infrastructure.Therefore, we propose to leverage complementary PUE metrics, coined xPUE, to compute the energy efficiency of the computing continuum from hardware components, up to the running software layers. Our contribution aims to deliver realtime energy efficiency metrics from different perspectives for cloud infrastructure, hence helping cloud ecosystems-from cloud providers to their customers-to experiment and optimize the energy usage of cloud infrastructures at large.

Paper Structure

This paper contains 29 sections, 9 equations, 12 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. pue
  4. DWPE
  5. $x$PUE: A New Family of Metrics
  6. Overview of $x$PUE
  7. : Assessing Cloud Servers
  8. : Assessing Cloud Services
  9. cpue: Applying $x$PUE Metrics to Cloud Infrastructures
  10. Revisiting State-of-the-Art Metrics with
  11. Implementation Details of $x$PUE Metrics
  12. Introducing the PowerAPI Toolkit
  13. Implementing the $x$PUE Formulas
  14. Deploying the $x$PUE Metrics
  15. Empirical Validation
  16. ...and 14 more sections

Figures (12)

  • Figure 1: Efficiency coverage of state-of-the-art and $x$PUE metrics (highlighted in grey and yellow, respectively).
  • Figure 2: Deployment of $x$PUE metrics using PowerAPI
  • Figure 3: Evolution of the over time and increasing workload
  • Figure 4: Correlation of the and the CPU average load
  • Figure 5: Comparing the of all the hardware configurations under test.
  • ...and 7 more figures