Student or educator? Sign up with your school email and get 6 months of Pro free.
CoChat

The DOI matches its Crossref record; 13 of the 13 references it cites were found in a registry or an index; 5 of the 16 sources its claims cite were checked for support.

Low-Capital Expenditure AI-Assisted Zero-Trust Control Plane for Brownfield Ethernet Environments

Open access
Wang, Hong-Sheng, Wang, Reen-Cheng · 2026 · 8th International Conference on Knowledge Innovation and Invention

Abstract

We developed an AI-assisted zero-trust control system at low capital expenditure to retrofit brownfield Ethernet environments without disruptive hardware upgrades or costly software-defined networking migration. Legacy network infrastructures in small and medium-sized enterprises (SMEs) lack the flexibility and programmability required by modern zero-trust architectures, creating a persistent security gap between static Layer-1 deployments and dynamic cyber threats. The developed system addresses this gap through a modular architecture that integrates genetic-algorithm-based virtual local area network (VLAN) optimization, large language model-guided firewall rule synthesis, threat-intelligence-driven policy automation, and telemetry-triggered adaptive isolation. Network assets are enumerated and evaluated through a risk-aware clustering model to enable micro-segmentation that aligns with the principle of least privilege. Optimized segmentation outputs are translated into pfSense firewall policies through structured prompt engineering and dual-stage validation, ensuring syntactic correctness and semantic consistency. A retrieval-augmented generation pipeline connects live telemetry with historical vulnerability intelligence, enabling rapid policy adjustments and automated containment responses. The system operates as an overlay on existing managed switches, orchestrating configuration changes through standards-compliant interfaces such as simple network management protocol and…

Overview

CoChat Summary

TL;DR

The authors built a modular, AI-assisted zero-trust control plane that overlays existing managed Ethernet switches to micro-segment brownfield SME networks without SDN hardware upgrades. On a 53-endpoint SME testbed it refined seven flat subnets into 34 purpose-specific VLANs and generated 851 validated pfSense firewall rules, raising ISO/IEC 27001 compliance from 62.3% to 94.7% and NIST CSF alignment from 58.9% to 91.2%.

Key findings

  • Segmentation granularity improved from 7 broad functional /24 subnets to 34 purpose-specific VLANs, with the GA converging at generation 120–150 and a final population diversity score of 0.160.
  • All 851 AI-generated pfSense rules passed Dual-Agent validation (100% acceptance), though 188 warnings were raised, mostly about missing logging on key services rather than syntax errors; the paper states over 90% of warnings were linked to enhanced logging.
  • Compliance scores rose: ISO/IEC 27001 from 62.3% to 94.7%, and NIST Cybersecurity Framework alignment from 58.9% to 91.2%.
  • The generated policy set prioritized core services (DNS, DHCP, NTP, LDAP), supported VoIP via SIP/RTP with bidirectional rules, restricted SNMP/SSH to authorized interfaces, and enforced default-deny for guest and IoT traffic.

Methods

  • Design: a four-module closed-loop control plane — (1) asset/rule-driven VLAN optimizer using a multidimensional pairwise risk function and GA with integer encoding, (2) GPT-4o firewall rule generation via two-tier prompt engineering and RAG, (3) daily threat-intelligence pipeline (NVD, CVE, MITRE CWE) with GPT-4o rule synthesis, and (4) telemetry-triggered adaptive isolation using a local vector database and RAG. Configuration is pushed over NETCONF/REST or SNMPv3.
  • GA parameters: population size 50, up to 500 generations, crossover probability 0.8, mutation probability 0.15, tournament selection, single-point crossover, heuristic mutation (40% group similar assets, 30% reduce VLAN count, 20% create new VLANs, 10% random); all cost-function weights set to 1.0.
  • Data/sample: an SME testbed of 53 endpoints (18 workstations across six departments, 4 mission-critical servers, 16 VoIP phones, 5 wireless access devices, 3 guest endpoints, 7 IoT sensors), initially seven /24 RFC1918 subnets on a core-access topology with one core switch and five access switches. Validation used a two-stage Dual-Agent mechanism (syntactic validation plus behavioral testing in a pfSense sandbox).

Limitations

  • Stated by the authors: the system still needs evaluation for long-term deployment stability, integration of adaptive learning for evolving threats, extension to hybrid cloud/container environments, and development of formal policy compliance verification.
  • Evident from the text: evaluation rests on a single 53-endpoint SME testbed with no comparison baseline or control condition, and the "low capital expenditure" claim is argued qualitatively — no cost figures, CapEx measurements, or ROI analysis are reported. The 188 validation warnings also indicate that logging/audit coverage gaps remained in the generated rule set.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.3390/engproc2025120054

Cites

13 of 13 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Verification

Claims the paper makes

3 Valid · 0 Not valid · 8 Not checked · 28 Own results and interpretations

13 more claims

Create a free account to see all 39 claims and to verify the 3 references still open.

See all claims — free account
  • Probably not a substantial claim. 10
  • It may state two separate claims as one. 14
  • The quote only partly states this claim. 2

Create a free account to see these claims.

Citations

Cites

Show all 10 references

Research works printed but not linked: 2

  1. Bellamkonda, 2022, Zero Trust Architecture Implementation: Strategies, Challenges, and Best Practices, Int. J. Commun. Netw. Inf. Secur., 14, 587
  2. Olaoye, G., Luz, A., John, O., and Akinleye, D. (2025, June 15). Network Segmentation and Micro-Segmentation Within Virtual Device Contexts. Available online: https://www.researchgate.net/publication/385788684_Network_Segmentation_and_Micro-Segmentation_within_Virtual_Device_Contexts.

Metadata

Authors
Wang, H.-S., & Wang, R.-C.
Affiliations
  • National Ilan University (TW)
  • National Taitung University (TW)
Corresponding author
Wang, Reen-Cheng
Venue
8th International Conference on Knowledge Innovation and Invention
Publisher
MDPI
Published
2026-02-05
Pages
54
OpenAlex
W7128073109
Cited by
1
References
10
Citations by year
2026: 1
FWCI
10.38
Type
conference paper
Language
en
Open access
Locations
  • IEEE International Conference on Knowledge Innovation and Invention (ICKII) (publishedVersion, cc-by)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion)
Topics
  • Physical Sciences › Computer Science › Computer Networks and Communications › Network Security and Intrusion Detection
  • Physical Sciences › Computer Science › Computer Networks and Communications › Software-Defined Networks and 5G
  • Physical Sciences › Computer Science › Computer Networks and Communications › Cybersecurity and Information Systems
Keywords
zero-trust security, Ethernet, genetic algorithm, large language models, prompt engineering, pfSense, micro-segmentation, retrieval-augmented generation, threat intelligence, principle of least privilege, small and medium-sized enterprises, network partitioning, NIST Cybersecurity Framework, ISO/IEC 27001, network telemetry, SNMP

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-10-08, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-08 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (gpt-6-sol).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Low-Capital Expenditure AI-Assisted Zero-Trust Control Plane for Brownfield Ethernet Environments. https://app.cochat.ai/papers/10.3390/engproc2025120054
The paper
Wang, H.-S., & Wang, R.-C. (2026). Low-Capital Expenditure AI-Assisted Zero-Trust Control Plane for Brownfield Ethernet Environments. 8th International Conference on Knowledge Innovation and Invention. https://doi.org/10.3390/engproc2025120054

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.