Skip to content

PNap: Lifecycle-aware Edge Multi-state sleep for Energy Efficient MEC

Home / Papers / PNap: Lifecycle-aware Edge Multi-state sleep for Energy Efficient MEC

PNap: Lifecycle-aware Edge Multi-state sleep for Energy Efficient MEC

Authors: Federico Giarrè, Holger Karl Date: 2026-03-24 Paper ID: arxiv:2603.23323

Summary

This paper addresses the high energy consumption in Multi-access Edge Computing (MEC) systems caused by keeping servers active during low-load periods. The authors propose PowerNap (PNap), a lifecycle-aware orchestration framework that jointly manages the server’s power state (multi-state sleep) and the service’s execution lifecycle. By integrating traffic forecasting, PNap determines the optimal balance between energy savings and service disruption avoidance. Evaluations show that PNap achieves up to 14.9% greater energy reduction than a state-of-the-art method while preserving comparable service availability metrics.

Key Contributions

  • Introduction of PowerNap (PNap), a framework for jointly managing server sleep states and service lifecycle states in MEC environments.
  • PNap leverages traffic forecasting to coordinate energy saving (server sleeping) with service continuity requirements (lifecycle operations).
  • Demonstration that PNap reduces energy consumption by up to 14.9% compared to a state-of-the-art solution while maintaining service availability.

Limitations

The paper focuses on MEC energy efficiency and assumes the availability of accurate traffic forecasting, whose implementation details or robustness limits are not deeply explored.

Open Questions & Future Work

Limitations

The paper focuses on MEC energy efficiency and assumes the availability of accurate traffic forecasting, whose implementation details or robustness limits are not deeply explored.

Metadata & Links

url
https://arxiv.org/abs/2603.23323
paper_id
2603.23323
paper_source
arxiv
domain
time-series
tags
forecastingenergy-efficienttime-seriesoptimizationdistributed-training
architectures
datasets
skill
TimeSeriesSkill
created_at
2026-03-25T21:18:17Z