Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

ORIENT: A Priority-Aware Energy-Efficient Approach for Latency-Sensitive Applications in 6G

Masoud Shokrnezhad, Tarik Taleb

TL;DR

This paper tackles the energy-efficient orchestration of 6G resources under latency-sensitive demands by formulating PIRA, a joint problem of service instance placement/assignment, path selection, and request prioritization. It proves NP-hardness and introduces ORIENT, a Graph Neural Network–assisted Double Dueling Deep Q-Learning framework to learn near-optimal policies that maximize profit while minimizing energy over time. The approach leverages an M/M/1 queuing model for latency, per-priority resource partitioning, and a GNN-based state representation to guide reinforcement learning. Results show ORIENT closely approaches the optimal solution and outperforms baselines under varying system sizes and request loads, highlighting practical potential for energy-aware, latency-conscious 6G orchestration.

Abstract

Anticipation for 6G's arrival comes with growing concerns about increased energy consumption in computing and networking. The expected surge in connected devices and resource-demanding applications presents unprecedented challenges for energy resources. While sustainable resource allocation strategies have been discussed in the past, these efforts have primarily focused on single-domain orchestration or ignored the unique requirements posed by 6G. To address this gap, we investigate the joint problem of service instance placement and assignment, path selection, and request prioritization, dubbed PIRA. The objective function is to maximize the system's overall profit as a function of the number of concurrently supported requests while simultaneously minimizing energy consumption over an extended period of time. In addition, end-to-end latency requirements and resource capacity constraints are considered for computing and networking resources, where queuing theory is utilized to estimate the Age of Information (AoI) for requests. After formulating the problem in a non-linear fashion, we prove its NP-hardness and propose a method, denoted ORIENT. This method is based on the Double Dueling Deep Q-Learning (D3QL) mechanism and leverages Graph Neural Networks (GNNs) for state encoding. Extensive numerical simulations demonstrate that ORIENT yields near-optimal solutions for varying system sizes and request counts.

ORIENT: A Priority-Aware Energy-Efficient Approach for Latency-Sensitive Applications in 6G

TL;DR

This paper tackles the energy-efficient orchestration of 6G resources under latency-sensitive demands by formulating PIRA, a joint problem of service instance placement/assignment, path selection, and request prioritization. It proves NP-hardness and introduces ORIENT, a Graph Neural Network–assisted Double Dueling Deep Q-Learning framework to learn near-optimal policies that maximize profit while minimizing energy over time. The approach leverages an M/M/1 queuing model for latency, per-priority resource partitioning, and a GNN-based state representation to guide reinforcement learning. Results show ORIENT closely approaches the optimal solution and outperforms baselines under varying system sizes and request loads, highlighting practical potential for energy-aware, latency-conscious 6G orchestration.

Abstract

Anticipation for 6G's arrival comes with growing concerns about increased energy consumption in computing and networking. The expected surge in connected devices and resource-demanding applications presents unprecedented challenges for energy resources. While sustainable resource allocation strategies have been discussed in the past, these efforts have primarily focused on single-domain orchestration or ignored the unique requirements posed by 6G. To address this gap, we investigate the joint problem of service instance placement and assignment, path selection, and request prioritization, dubbed PIRA. The objective function is to maximize the system's overall profit as a function of the number of concurrently supported requests while simultaneously minimizing energy consumption over an extended period of time. In addition, end-to-end latency requirements and resource capacity constraints are considered for computing and networking resources, where queuing theory is utilized to estimate the Age of Information (AoI) for requests. After formulating the problem in a non-linear fashion, we prove its NP-hardness and propose a method, denoted ORIENT. This method is based on the Double Dueling Deep Q-Learning (D3QL) mechanism and leverages Graph Neural Networks (GNNs) for state encoding. Extensive numerical simulations demonstrate that ORIENT yields near-optimal solutions for varying system sizes and request counts.
Paper Structure (18 sections, 13 equations, 2 figures, 1 table, 1 algorithm)

This paper contains 18 sections, 13 equations, 2 figures, 1 table, 1 algorithm.

Table of Contents

  1. Introduction
  2. System Model
  3. Resources
  4. Requests
  5. Problem Definition
  6. Instance Orchestration Constraints
  7. Path Selection Constraints
  8. Request Prioritization Constraints
  9. Latency Constraints
  10. Objective Function
  11. Problem
  12. Complexity Analysis
  13. ORIENT
  14. Learning Mechanism
  15. Agent Customization
  16. ...and 3 more sections

Figures (2)

  • Figure 1: The system model, including network devices and distributed compute nodes facilitating holographic telepresence services for end users.
  • Figure 2: The mean energy consumption of supported requests and the total profit vs. the system size (A & B) and the number of all requests (C & D).