pith. sign in

arxiv: 2604.24181 · v1 · submitted 2026-04-27 · 💻 cs.NI

Optimizing power by selective IP card shutdown using transport slicing

Alfonso S\'anchez-Maci\'an , \'Oscar Gonz\'alez de Dios , Jos\'e Alberto Hern\'andez , Liesbeth Roelens , Pablo Armingol Robles , Juan Pedro Fernadez-Palacios , Ram\'on Casellas , Filippo Cugini This is my paper

Pith reviewed 2026-05-08 01:17 UTC · model grok-4.3

classification 💻 cs.NI
keywords IP network slicingenergy efficiencyline card shutdownmulti-objective optimizationtransport slicinglatency constraintsnight slice6G power savings
0
0 comments X

The pith

A dual-slice IP strategy lets operators deactivate more than 40 percent of line cards during low-traffic periods by switching between a performance-focused daytime slice and an energy-minimized nighttime slice.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper proposes splitting an IP network into two transport slices: a Day Slice sized for peak loads and a Night Slice that turns off excess line cards to save power. Pareto-based evolutionary algorithms balance the resulting energy use against end-to-end latency. Experiments on the SNDlib india35 topology show the Night Slice can shut down over 40 percent of cards while latency stays acceptable. A multi-service version using AGE-MOEA further enforces stricter latency for premium flows without losing most of the energy gains.

Core claim

By defining separate Day and Night slices and switching traffic between them at fixed times, an IP router network can selectively power down line cards in the Night Slice during off-peak hours; multi-objective optimization with NSGA-II, CTAEA, and AGE-MOEA keeps latency within bounds, yielding more than 40 percent card deactivation on the india35 topology and preserving substantial savings even when premium traffic receives differentiated QoS limits below 7 ms.

What carries the argument

The dual-slice transport mechanism that routes traffic to a reduced-card Night Slice at predefined low-demand intervals, with Pareto evolutionary algorithms jointly minimizing energy and latency.

If this is right

  • More than 40 percent of line cards can remain off during low-traffic intervals while latency stays within operator targets.
  • Multi-objective evolutionary search finds practical trade-offs between energy reduction and delay on realistic topologies.
  • Premium traffic can be given tighter latency guarantees in the same Night Slice without erasing most of the energy benefit.
  • The approach applies directly to existing card-based IP routers without requiring new hardware.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Dynamic rather than fixed-time slice switching could further increase savings if traffic prediction improves.
  • The same slicing pattern might extend to other router types or optical layers if line-card power models are updated.
  • Operators facing rising 6G and AI traffic could combine this method with existing energy-aware routing protocols.

Load-bearing premise

Traffic can be moved between the Day and Night slices at set times without breaking sessions or causing large latency spikes, provided existing traffic engineering handles the hand-off.

What would settle it

Live measurements on the india35 topology (or an equivalent network) showing that slice switches produce latency jumps above the modeled bounds or cause session drops would falsify the claimed energy savings.

Figures

Figures reproduced from arXiv: 2604.24181 by Alfonso S\'anchez-Maci\'an, Filippo Cugini, Jos\'e Alberto Hern\'andez, Juan Pedro Fernadez-Palacios, Liesbeth Roelens, \'Oscar Gonz\'alez de Dios, Pablo Armingol Robles, Ram\'on Casellas.

Figure 1
Figure 1. Figure 1: Division of energy consumption by component in a typical router line card. view at source ↗
Figure 2
Figure 2. Figure 2: Average and maximum power consumption for 100G, 40G, and 10G cards as a function of the number of view at source ↗
Figure 3
Figure 3. Figure 3: India35 topology from SNDLib[22] DE algorithm maintains a population of candidate solutions (sets of weights), which evolve over several generations through three steps: mutation, recombination, and selection. First, a population (pool) of different random sets of link weights (x1...xp) is created. The total number of elements p in this first population is calculated as p = popsize ∗ N, where popsize is a … view at source ↗
Figure 4
Figure 4. Figure 4: Links with active ports for the DE solution view at source ↗
Figure 5
Figure 5. Figure 5: Links with active ports for the DA solution view at source ↗
Figure 6
Figure 6. Figure 6: Path automatically generated by Networkx view at source ↗
Figure 7
Figure 7. Figure 7: Traffic per link in different scenarios 4.2 Multi-objective optimization The multi-objective optimization experiments also use the india35 topology. Its moderate size (35 nodes, 80 links) allows exhaustive traffic simulation and Pareto-based optimization without excessive computational cost. Same configuration is used in terms of capacity of the ports, line card configurations and daytime and nighttime tra… view at source ↗
Figure 8
Figure 8. Figure 8: Pareto front obtained with the NSGA-II algorithm. view at source ↗
Figure 9
Figure 9. Figure 9: Pareto front obtained with the CTAEA algorithm. view at source ↗
Figure 10
Figure 10. Figure 10: Pareto front obtained with the AGE-MOEA algorithm. view at source ↗
Figure 11
Figure 11. Figure 11: Active paths for the best solution in terms of cards. view at source ↗
Figure 12
Figure 12. Figure 12: Pareto fronts (cards vs average premium latency) without latency constraint. view at source ↗
Figure 13
Figure 13. Figure 13: Pareto fronts (cards vs max premium latency) without latency constraint. view at source ↗
Figure 14
Figure 14. Figure 14: Pareto fronts (cards vs average premium latency) with latency constraint. view at source ↗
Figure 15
Figure 15. Figure 15: Pareto fronts (cards vs max premium latency) without latency constraint. view at source ↗
read the original abstract

The increasing energy demands of upcoming sixth-generation (6G) mobile networks and networks supporting AI applications pose significant challenges for network operators in terms of operational costs and environmental impact. To address these challenges, this paper proposes a novel IP-based network slicing strategy that optimizes energy efficiency through a dual-slice approach. The proposed solution consists of a Day Slice, designed to meet high-performance requirements during peak traffic hours, and a Night Slice, optimized for energy savings by deactivating excess line-cards in card-based routers during periods of low traffic demand. The traffic is switched between the Day and Night Slices at predefined times, assuming appropriate traffic engineering mechanisms are in place to minimize disruption and support session continuity. We apply Pareto-based evolutionary algorithms (NSGA-II, CTAEA, and AGE-MOEA) to jointly optimize energy consumption and latency. Experiments conducted on the SNDlib india35 topology demonstrate that multi-objective optimization can deactivate over 40% of line cards during low-traffic periods, providing significant energy savings while maintaining acceptable performance. Additionally, a multi-service extension using AGE-MOEA introduces differentiated QoS constraints, maintaining latency below 7 ms for premium traffic while preserving substantial energy savings.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The paper proposes a dual-slice IP transport strategy for energy-efficient 6G networks: a Day Slice sized for peak traffic and a Night Slice that deactivates excess line cards in card-based routers during low-demand periods. Traffic is switched between slices at fixed times under the assumption that existing traffic-engineering mechanisms ensure session continuity and acceptable performance. Multi-objective evolutionary algorithms (NSGA-II, CTAEA, AGE-MOEA) are used to jointly minimize energy and latency; on the SNDlib india35 topology the approach reports deactivation of over 40% of line cards while keeping latency acceptable, with a multi-service extension that enforces differentiated QoS.

Significance. If the unmodeled switching assumption can be substantiated, the work offers a concrete, topology-grounded demonstration that standard Pareto-based MOEAs can identify substantial line-card shutdown opportunities in IP networks without violating latency targets. The use of a public SNDlib topology and three distinct algorithms is a positive reproducibility feature.

major comments (2)
  1. [Abstract and experimental evaluation] The headline result (over 40% line-card deactivation on india35) is obtained by optimizing two independent static routing problems (Day and Night slices). No model of the slice-transition process, session hand-off, routing convergence, or transient latency/loss appears in the formulation or experiments, yet the practical energy-saving claim is meaningful only if this transition can be performed with “acceptable performance.” This assumption is load-bearing for the central contribution.
  2. [Experimental evaluation] The experimental section supplies no traffic matrix details, exact algorithm parameter settings (population size, generations, crossover/mutation rates), error bars, or comparison against single-objective or heuristic baselines. Without these, the 40% deactivation figure cannot be independently verified or assessed for robustness.
minor comments (1)
  1. [Multi-service extension] The multi-service extension using AGE-MOEA is mentioned only briefly; clarifying how the premium-traffic latency constraint (<7 ms) is encoded in the objective/constraint set would improve reproducibility.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback and positive evaluation of the paper's significance and reproducibility features. We address each major comment below and will incorporate revisions to strengthen the manuscript.

read point-by-point responses

  1. Referee: [Abstract and experimental evaluation] The headline result (over 40% line-card deactivation on india35) is obtained by optimizing two independent static routing problems (Day and Night slices). No model of the slice-transition process, session hand-off, routing convergence, or transient latency/loss appears in the formulation or experiments, yet the practical energy-saving claim is meaningful only if this transition can be performed with “acceptable performance.” This assumption is load-bearing for the central contribution.

    Authors: We agree that the transition process is not modeled and that the assumption regarding traffic-engineering mechanisms for session continuity is central to the practical claim. The manuscript explicitly states this assumption in the abstract and introduction but does not simulate transients. In the revised version, we will add a new subsection in the methodology discussing the assumption, citing relevant literature on IP/MPLS traffic engineering and SDN-based handover techniques that support low-disruption slice switching. We will also add a limitations paragraph noting that transient latency and loss are not evaluated in the current static optimization framework. This provides substantiation via references without expanding the scope to full dynamic simulation. revision: partial

  2. Referee: [Experimental evaluation] The experimental section supplies no traffic matrix details, exact algorithm parameter settings (population size, generations, crossover/mutation rates), error bars, or comparison against single-objective or heuristic baselines. Without these, the 40% deactivation figure cannot be independently verified or assessed for robustness.

    Authors: We agree that these omissions hinder independent verification. In the revised manuscript, we will: (1) specify the traffic matrix details or generation procedure for the india35 topology, (2) report exact parameter settings for NSGA-II, CTAEA, and AGE-MOEA including population size, generations, crossover and mutation rates, (3) include error bars from multiple independent runs of each algorithm, and (4) add comparisons to a single-objective genetic algorithm baseline and a simple heuristic (e.g., utilization-threshold-based card deactivation). These additions will be placed in the experimental evaluation section. revision: yes

Circularity Check

0 steps flagged

No circularity: optimization results derive from standard algorithms on public topology

full rationale

The paper applies off-the-shelf Pareto evolutionary algorithms (NSGA-II, CTAEA, AGE-MOEA) to jointly minimize energy and latency on two static slices defined over the public SNDlib india35 topology. The reported >40% line-card deactivation is the direct numerical output of those runs under the stated traffic matrices; no parameter is fitted to the target metric and then re-labeled as a prediction, no self-citation supplies a uniqueness theorem, and the slice-switching assumption is declared as an external prerequisite rather than being embedded in the optimization equations. The derivation chain is therefore self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

2 free parameters · 1 axioms · 0 invented entities

The central claim rests on domain assumptions about predictable traffic and smooth slice switching rather than new physical entities or many fitted constants; the optimization itself introduces no additional free parameters beyond standard algorithm settings.

free parameters (2)
  • Slice switching times
    Predefined times chosen to align with assumed traffic demand patterns; exact values not reported in abstract.
  • Latency threshold for premium traffic
    Set at 7 ms in the multi-service extension; chosen to meet QoS targets.
axioms (1)
  • domain assumption Appropriate traffic engineering mechanisms exist that can switch slices at fixed times while minimizing disruption and preserving session continuity
    Explicitly stated as an assumption required for the dual-slice approach to work without service impact.

pith-pipeline@v0.9.0 · 5548 in / 1452 out tokens · 92629 ms · 2026-05-08T01:17:55.191870+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

22 extracted references · 22 canonical work pages

  1. [1]

    A framework for network slices in networks built from ietf technologies

    Adrian Farrel, John Drake, Reza Rokui, Shunsuke Homma, Kiran Makhijani, Luis M Contreras, and Jeff Tantsura. A framework for network slices in networks built from ietf technologies. Technical report, RFC 9543, IETF, Tech. 17 APREPRINT- APRIL28, 2026 Rep. 9543, 2024

    work page 2026

  2. [2]

    Efficient network slicing with sdn and heuristic algorithm for low latency services in 5g/b5g networks.Sensors, 23(13), 2023

    Robert Botez, Andres-Gabriel Pasca, Alin-Tudor Sferle, Iustin-Alexandru Ivanciu, and Virgil Dobrota. Efficient network slicing with sdn and heuristic algorithm for low latency services in 5g/b5g networks.Sensors, 23(13), 2023

    work page 2023

  3. [3]

    A comprehensive overview of network slicing for improving the energy efficiency of fifth-generation networks.Sensors, 24(10), 2024

    Josip Lorincz, Amar Kukuruzovi ´c, and Zoran Blaževi ´c. A comprehensive overview of network slicing for improving the energy efficiency of fifth-generation networks.Sensors, 24(10), 2024

    work page 2024

  4. [4]

    Energy saving technologies and best practices for 5g radio access network.IEEE Access, 10:51747–51756, 2022

    Rumeng Tan, Ying Shi, Yingying Fan, Wentao Zhu, and Tong Wu. Energy saving technologies and best practices for 5g radio access network.IEEE Access, 10:51747–51756, 2022

    work page 2022

  5. [5]

    Energy-optimal end-to-end network slicing in cloud-based architecture.IEEE Open Journal of the Communications Society, 3:574–592, 2022

    Meysam Masoudi, Özlem Tu˘gfe Demir, Jens Zander, and Cicek Cavdar. Energy-optimal end-to-end network slicing in cloud-based architecture.IEEE Open Journal of the Communications Society, 3:574–592, 2022

    work page 2022

  6. [6]

    Jorswieck

    Bho Matthiesen, Osman Aydin, and Eduard A. Jorswieck. Throughput and energy-efficient network slicing. In WSA 2018; 22nd International ITG Workshop on Smart Antennas, pages 1–6, 2018

    work page 2018

  7. [7]

    Time efficient energy- aware routing in software defined networks

    Yu-Hao Chen, Tai-Lin Chin, Chin-Ya Huang, Shan-Hsiang Shen, and Ruei-Yan Huang. Time efficient energy- aware routing in software defined networks. In2018 IEEE 7th International Conference on Cloud Networking (CloudNet), pages 1–7, 2018

    work page 2018

  8. [8]

    Energy-aware routing in sdn enabled data center network

    Priyanka Kamboj and Sujata Pal. Energy-aware routing in sdn enabled data center network. In2022 IEEE 21st International Symposium on Network Computing and Applications (NCA), volume 21, pages 123–130, 2022

    work page 2022

  9. [9]

    Multi-stage power- aware intelligent adaptive routing algorithms in bundled links based backbone networks.IEEE Access, 10:109863– 109893, 2022

    Jinhong Zhang, Xingwei Wang, Xiancun Zhou, Min Huang, Ruixia Li, and Zhonghua Wu. Multi-stage power- aware intelligent adaptive routing algorithms in bundled links based backbone networks.IEEE Access, 10:109863– 109893, 2022

    work page 2022

  10. [10]

    Tepg: a traffic engineering based power-aware greedy routing algorithm in backbone networks with bundled links.Frontiers of Computer Science, 18:183313, 2024

    Jinhong Zhang, Xingwei Wang, Ruixia Li, Bo Yi, Min Huang, and Dongxing Shui. Tepg: a traffic engineering based power-aware greedy routing algorithm in backbone networks with bundled links.Frontiers of Computer Science, 18:183313, 2024

    work page 2024

  11. [11]

    Network energy-saving adjustment routing under changing demands: Models and algorithms.IEEE Access, 8:90676–90685, 2020

    Mengjia Zhang, Wenguo Yang, Suixiang Gao, and Wei Xu. Network energy-saving adjustment routing under changing demands: Models and algorithms.IEEE Access, 8:90676–90685, 2020

    work page 2020

  12. [12]

    Does rate adaptation at daily timescales make sense? In Proceedings of the 2nd Workshop on Sustainable Computer Systems, HotCarbon ’23, New York, NY , USA, 2023

    Romain Jacob, Jackie Lim, and Laurent Vanbever. Does rate adaptation at daily timescales make sense? In Proceedings of the 2nd Workshop on Sustainable Computer Systems, HotCarbon ’23, New York, NY , USA, 2023. Association for Computing Machinery

    work page 2023

  13. [13]

    Saving energy in ip-over-wdm networks by switching off line cards in low-demand scenarios

    Filip Idzikowski, Sebastian Orlowski, Christian Raack, Hagen Woesner, and Adam Wolisz. Saving energy in ip-over-wdm networks by switching off line cards in low-demand scenarios. In2010 14th Conference on Optical Network Design and Modeling (ONDM), pages 1–6, 2010

    work page 2010

  14. [14]

    Q-learning and simulated annealing-based routing for software-defined networks

    Marwa Kandil, Mohamad Khattar Awad, Eiman Mohammed Alotaibi, and Reza Mohammadi. Q-learning and simulated annealing-based routing for software-defined networks. In2022 International Conference on Computer and Applications (ICCA), pages 1–10, 2022

    work page 2022

  15. [15]

    Energy-saving method in sdn networks based on deployment location and link traffic

    Hangyu Wang. Energy-saving method in sdn networks based on deployment location and link traffic. In2023 3rd International Symposium on Computer Technology and Information Science (ISCTIS), pages 69–74, 2023

    work page 2023

  16. [16]

    Lessons learned from ip routers power measurements and characterization

    David de la Osa Mostazo, Pablo Armingol Robles, Óscar González de Dios, and Juan Pedro Fernández- Palacios Giménez. Lessons learned from ip routers power measurements and characterization. In2024 15th International Conference on Network of the Future (NoF), pages 245–253, 2024

    work page 2024

  17. [17]

    Springer Berlin Heidelberg, Berlin, Heidelberg, 2004

    Jouni Lampinen and Rainer Storn.Differential Evolution, pages 123–166. Springer Berlin Heidelberg, Berlin, Heidelberg, 2004

    work page 2004

  18. [18]

    Generalized simulated annealing algorithm and its application to the thomson model.Physics Letters A, 233(3):216–220, 1997

    Y Xiang, D.Y Sun, W Fan, and X.G Gong. Generalized simulated annealing algorithm and its application to the thomson model.Physics Letters A, 233(3):216–220, 1997

    work page 1997

  19. [19]

    K. Deb, A. Pratap, S. Agarwal, and T. Meyarivan. A fast and elitist multiobjective genetic algorithm: Nsga-ii. IEEE Transactions on Evolutionary Computation, 6(2):182–197, 2002

    work page 2002

  20. [20]

    Two-archive evolutionary algorithm for constrained multiobjective optimization.IEEE Transactions on Evolutionary Computation, 23(2):303–315, 2019

    Ke Li, Renzhi Chen, Guangtao Fu, and Xin Yao. Two-archive evolutionary algorithm for constrained multiobjective optimization.IEEE Transactions on Evolutionary Computation, 23(2):303–315, 2019

    work page 2019

  21. [21]

    An adaptive evolutionary algorithm based on non-euclidean geometry for many-objective optimization

    Annibale Panichella. An adaptive evolutionary algorithm based on non-euclidean geometry for many-objective optimization. InProceedings of the Genetic and Evolutionary Computation Conference, GECCO ’19, page 595–603, New York, NY , USA, 2019. Association for Computing Machinery

    work page 2019

  22. [22]

    Orlowski, R

    S. Orlowski, R. Wessäly, M. Pióro, and A. Tomaszewski. Sndlib 1.0—survivable network design library.Networks, 55(3):276–286, 2010. 18

    work page 2010