Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Energy Metrics for Edge Microservice Request Placement Strategies

Klervie Toczé, Simin Nadjm-Tehrani

TL;DR

This work addresses energy-centric microservice request placement at the edge, examining whether different energy optimization goals yield distinct placements. It formulates two energy metrics, $E^O$ (overall energy) and $E^M$ (marginal energy), within a Traveling Purchaser Problem framework and solves the resulting ILP to study placement decisions under latency constraints. The results show that the two metrics indeed lead to different placements, particularly under heterogeneous load and larger function-instance spaces, underscoring the importance of metric selection in energy-aware edge scheduling. The findings motivate exploring adaptive or two-tier placement strategies to balance energy efficiency with timely service delivery in real-world, larger-scale edge deployments.

Abstract

Microservices are a way of splitting the logic of an application into small blocks that can be run on different computing units and used by other applications. It has been successful for cloud applications and is now increasingly used for edge applications. This new architecture brings many benefits but it makes deciding where a given service request should be executed (i.e. its placement) more complex as every small block needed for the request has to be placed. In this paper, we investigate energy-centric request placement for services that use the microservice architecture, and specifically whether using different energy metrics for optimization leads to different placement strategies. We consider the problem as an instance of a traveling purchaser problem and propose an integer linear programming formulation. This formulation aims at minimizing energy consumption while respecting latency requirements. We consider two different energy consumption metrics, namely overall or marginal energy, when applied as a measure to determine a placement. Our simulations show that using different energy metrics indeed results in different request placements. The paper presents several parameters influencing the extent of this difference.

Energy Metrics for Edge Microservice Request Placement Strategies

TL;DR

This work addresses energy-centric microservice request placement at the edge, examining whether different energy optimization goals yield distinct placements. It formulates two energy metrics, (overall energy) and (marginal energy), within a Traveling Purchaser Problem framework and solves the resulting ILP to study placement decisions under latency constraints. The results show that the two metrics indeed lead to different placements, particularly under heterogeneous load and larger function-instance spaces, underscoring the importance of metric selection in energy-aware edge scheduling. The findings motivate exploring adaptive or two-tier placement strategies to balance energy efficiency with timely service delivery in real-world, larger-scale edge deployments.

Abstract

Microservices are a way of splitting the logic of an application into small blocks that can be run on different computing units and used by other applications. It has been successful for cloud applications and is now increasingly used for edge applications. This new architecture brings many benefits but it makes deciding where a given service request should be executed (i.e. its placement) more complex as every small block needed for the request has to be placed. In this paper, we investigate energy-centric request placement for services that use the microservice architecture, and specifically whether using different energy metrics for optimization leads to different placement strategies. We consider the problem as an instance of a traveling purchaser problem and propose an integer linear programming formulation. This formulation aims at minimizing energy consumption while respecting latency requirements. We consider two different energy consumption metrics, namely overall or marginal energy, when applied as a measure to determine a placement. Our simulations show that using different energy metrics indeed results in different request placements. The paper presents several parameters influencing the extent of this difference.

Paper Structure

This paper contains 43 sections, 25 equations, 15 figures, 7 tables.

Table of Contents

  1. Introduction
  2. System model
  3. Edge infrastructure
  4. Service
  5. Service request
  6. Request placement
  7. Request completion time
  8. Energy-centric placement
  9. Energy models
  10. Edge links
  11. Edge devices
  12. Energy metrics
  13. Associated envisioned placement strategies
  14. Overall energy
  15. Marginal energy
  16. ...and 28 more sections

Figures (15)

  • Figure 1: Categorization of the placements for a given request with varying load.
  • Figure 2: Power profile of a Parasilo device according to Ahvar_EstimatingEnergy
  • Figure 3: Categorization of the placements obtained for a given request with different edge device loads.
  • Figure 4: Categorization of the placements obtained for a given request with varying load and function instance availability.
  • Figure 5: Average included device utilization with varying function instance availability and infrastructure load levels.
  • ...and 10 more figures