Evaluasi Komparatif dan Benchmarking Eksperimental Efisiensi Multi-Dimensi WireGuard, IPSec, dan OpenVPN pada Virtualisasi KVM Linux Kernel 6.8
Open accessThe DOI matches its Crossref record; 22 of the 22 references it cites were found in a registry or an index; none of the 5 sources its claims cite is checked for support yet.
Abstract
Peningkatan penggunaan layanan cloud saat ini memaksa pembaruan pada standar keamanan jaringan yang mengutamakan efisiensi pemanfaatan sumber daya. Namun, literatur saat ini sering kali terpaku pada evaluasi throughput secara parsial tanpa mempertimbangkan optimasi arsitektural pada kernel modern. Evaluasi mendalam terhadap WireGuard, IPSec, dan OpenVPN dilakukan pada ekosistem Linux Kernel 6.8+ melalui desain eksperimental multi-fase dengan 30 iterasi (n=30) menggunakan tool iperf3. Validitas data diuji secara ketat melalui ANOVA satu arah dan uji post-hoc Bonferroni guna memastikan signifikansi perbedaan performa. Data pengujian menunjukkan dominasi WireGuard dengan rata-rata throughput 842,79 Mbps (p<0,001). Protokol ini juga mencatatkan skor efisiensi tertinggi sebesar 97,6/100, yang dihitung berdasarkan pembobotan komposit antara throughput, latensi, dan beban kerja CPU. Kebaruan penelitian ini terletak pada pengenalan Indeks Efisiensi Multi-Dimensi dan model skalabilitas prediktif sebagai alternatif atas metode benchmarking tradisional yang bersifat fragmentatif. Hasil ini memberikan kerangka kerja strategis bagi profesional TI untuk memproyeksikan performa jaringan pada infrastruktur berskala masif dengan landasan data yang lebih presisi.
Overview
CoChat Summary
TL;DR
This paper benchmarks WireGuard, IPSec, and OpenVPN on Ubuntu 24.04.3 LTS with Linux Kernel 6.8 under KVM virtualization, using 30 iterations and iperf3. WireGuard shows the highest throughput and a 97.6/100 multi-dimensional efficiency score, with statistically significant advantages over IPSec and OpenVPN.
Key findings
- WireGuard averaged 842.79 Mbps (SD 41.94; 95% CI 827.78–857.80), versus IPSec 477.16 Mbps (SD 36.65; CI 464.05–490.28) and OpenVPN 316.82 Mbps (SD 27.56; CI 306.96–326.68). One-way ANOVA gave F = 1693.50, p < 0.000001, η² = 0.87, with Bonferroni post-hoc confirming significant pairwise differences.
- Latency: WireGuard RTT 60.81 ms, OpenVPN 77.75 ms, IPSec 84.77 ms; packet loss was 0% for all three. Table 1 reports memory usage of 3,965 MiB for WireGuard, 2,793 MiB for OpenVPN, and 5,789 MiB for IPSec.
- Scalability projections: WireGuard max practical throughput 520 Mbps, max concurrent connections 80, memory per connection 0.8 MB; IPSec 380 Mbps / 45 / 1.2 MB; OpenVPN 250 Mbps / 25 / 1.5 MB. WireGuard reached 80% CPU at 80 connections, while OpenVPN saturated 100% CPU at 25 connections. In two-way stress testing, WireGuard maintained 52.0 Gbps aggregate throughput.
- The paper reports a WireGuard efficiency score of 97.6/100, calculated from a composite weighting of throughput, latency, and CPU load.
Methods
- Controlled KVM virtualization testbed on one high-performance mobile workstation: Ubuntu 24.04.3 LTS (Linux 6.8.0-100-generic), AMD Ryzen 3 5300U (4 cores/8 threads), 20 GB DDR4, AES-NI/SHA/AVX2, NVMe SSD, Realtek NIC; two virtual nodes (client and server) connected via a KVM virtual switch at 10 Gbps, with CPU pinning to reduce jitter.
- Protocol configurations: WireGuard with ChaCha20-Poly1305 in kernel-space, MTU 1420; IPSec in tunnel mode using StrongSwan IKEv2 with AES-256-GCM and AES-NI, MTU 1440; OpenVPN in UDP mode with Data Channel Offload (DCO) and AES-256-GCM, MTU 1500.
- Five-phase procedure: core performance, main performance, stress/reliability (including 24-hour endurance), post-test statistical analysis, and reporting/visualization. Testing used 30 iterations per protocol, iperf3 for 60 s with alternating TCP/UDP, a 10 s warm-up, 900 data points per protocol, Shapiro-Wilk and Levene’s tests, one-way ANOVA, Bonferroni post-hoc, and multivariate linear regression for scalability projection.
Limitations
- Stated: results are tied to AMD Ryzen infrastructure with AVX2 support and Linux Kernel 6.8+; generalization to different OS memory management is limited; the efficiency score and scalability model are an initial foundation; future work is needed for post-quantum cryptography and ARM architectures.
- Evident from the text: experiments used a single physical workstation with a quad-core Ryzen 3 5300U, so hardware-specific findings may not generalize. The paper also reports inconsistent throughput values between Table 1 (e.g., WireGuard 30,494.31) and Table 2/abstract (842.79 Mbps), which complicates direct interpretation.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Figures









Verification
Claims the paper makes
0 Valid · 2 Not valid · 3 Not checked · 42 Own results and interpretations
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Pola ini sangat kontras dengan OpenVPN yang menunjukkan fluktuasi permintaan memori yang lebih tinggi seiring dengan kompleksitas manajemen sesi SSL/TLS di dalam memori sistem sebagaimana ditunjukkan pada Gambar 4.”
§ Evaluasi Komparatif dan Benchmarking Eksperimental Efisiensi Multi-Dimensi WireGuard, IPSec, dan OpenVPN pada Virtualisasi KVM Linux Kernel > 3.2.3. Korelasi Penggunaan Memori terhadap Skalabilitas
boostedstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“uji normalitas menggunakan Shapiro-Wilk Test menunjukkan bahwa distribusi data throughput pada setiap grup protokol mengikuti distribusi normal”
§ Evaluasi Komparatif dan Benchmarking Eksperimental Efisiensi Multi-Dimensi WireGuard, IPSec, dan OpenVPN pada Virtualisasi KVM Linux Kernel > 3.3.1. Verifikasi Asumsi Statistik Dasar
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Sebaliknya, IPSec mencatatkan latensi tertinggi sebesar 84,77 ms.”
§ Evaluasi Komparatif dan Benchmarking Eksperimental Efisiensi Multi-Dimensi WireGuard, IPSec, dan OpenVPN pada Virtualisasi KVM Linux Kernel > 3.1. Evaluasi Performa Inti
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Fenomena ini divalidasi oleh koefisien korelasi negatif yang sangat kuat antara throughput dan penggunaan CPU, yang membuktikan bahwa efisiensi kode WireGuard memungkinkan pemrosesan paket data yang lebih besar dengan beban komputasi minimal.”
§ Evaluasi Komparatif dan Benchmarking Eksperimental Efisiensi Multi-Dimensi WireGuard, IPSec, dan OpenVPN pada Virtualisasi KVM Linux Kernel > 3.1. Evaluasi Performa Inti
boostedstates an association- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Hasil uji menunjukkan bahwa WireGuard secara konsisten berada di luar interval confidence kedua protokol lainnya, yang mengonfirmasi posisinya sebagai protokol dengan performa paling stabil dan efisien pada sistem berbasis Ryzen 3.”
§ Evaluasi Komparatif dan Benchmarking Eksperimental Efisiensi Multi-Dimensi WireGuard, IPSec, dan OpenVPN pada Virtualisasi KVM Linux Kernel > 3.3.4. Uji Post-Hoc Bonferroni
boosted- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
16 more claims
Create a free account to see all 47 claims and to verify the 2 references still open.
- It may state two separate claims as one. 14
- The quote only partly states this claim. 3
- Probably not a substantial claim. 11
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Rated major: it would touch a result — a value a result rests on, reported and recomputed p-values on opposite sides of the significance level, two passages that conflict with no sign which is right, or a cited work retracted before this paper was published.
In the paper
§ Evaluasi Komparatif dan Benchmarking Eksperimental Efisiensi Multi-Dimensi WireGuard, IPSec, dan OpenVPN pada Virtualisasi KVM Linux Kernel > 3.1. Evaluasi Performa Inti
“Selama pengujian tekanan dua arah, WireGuard mempertahankan throughput agregat 52,0 Gbps dengan variansi latency terendah”
§ Evaluasi Komparatif dan Benchmarking Eksperimental Efisiensi Multi-Dimensi WireGuard, IPSec, dan OpenVPN pada Virtualisasi KVM Linux Kernel > 2.3.2. Konfigurasi Bridge Networking 10Gbps
“Antarmuka jembatan virtual (br0) dibuat untuk menghubungkan Client Node dan Server Node pada kecepatan teoritis hingga 10Gbps.”
Rated major: it would touch a result — a value a result rests on, reported and recomputed p-values on opposite sides of the significance level, two passages that conflict with no sign which is right, or a cited work retracted before this paper was published.
In the paper
§ Evaluasi Komparatif dan Benchmarking Eksperimental Efisiensi Multi-Dimensi WireGuard, IPSec, dan OpenVPN pada Virtualisasi KVM Linux Kernel > 3.1. Evaluasi Performa Inti
“Tabel 1. Perbandingan Metrik Performa Inti Protokol Throughput (Mbps) Latensi Ratarata (ms) Packet Loss (%) Penggunaan Memori (MiB) WireGuard 30.494,31 60,81 0 3,965”
§ Evaluasi Komparatif dan Benchmarking Eksperimental Efisiensi Multi-Dimensi WireGuard, IPSec, dan OpenVPN pada Virtualisasi KVM Linux Kernel > 2.3.2. Konfigurasi Bridge Networking 10Gbps
“Antarmuka jembatan virtual (br0) dibuat untuk menghubungkan Client Node dan Server Node pada kecepatan teoritis hingga 10Gbps.”
Rated major: it would touch a result — a value a result rests on, reported and recomputed p-values on opposite sides of the significance level, two passages that conflict with no sign which is right, or a cited work retracted before this paper was published.
In the paper
§ Evaluasi Komparatif dan Benchmarking Eksperimental Efisiensi Multi-Dimensi WireGuard, IPSec, dan OpenVPN pada Virtualisasi KVM Linux Kernel > 3.1. Evaluasi Performa Inti
“WireGuard 30.494,31 60,81 0 3,965 OpenVPN 29.284,70 77,75 0 2,793 IPSec 28.475,54 84,77 0 5,789”
§ Evaluasi Komparatif dan Benchmarking Eksperimental Efisiensi Multi-Dimensi WireGuard, IPSec, dan OpenVPN pada Virtualisasi KVM Linux Kernel > 3.1. Evaluasi Performa Inti
“IPSec 477,16 36,65 464,05 -490,28 OpenVPN 316,82 27,56 306,96 -326,68”
Rated major: it would touch a result — a value a result rests on, reported and recomputed p-values on opposite sides of the significance level, two passages that conflict with no sign which is right, or a cited work retracted before this paper was published.
In the paper
§ Evaluasi Komparatif dan Benchmarking Eksperimental Efisiensi Multi-Dimensi WireGuard, IPSec, dan OpenVPN pada Virtualisasi KVM Linux Kernel > 3.1. Evaluasi Performa Inti
“WireGuard 842,79 41,94 827,78 -857,80”
§ Evaluasi Komparatif dan Benchmarking Eksperimental Efisiensi Multi-Dimensi WireGuard, IPSec, dan OpenVPN pada Virtualisasi KVM Linux Kernel > 3.2.2. Analisis Saturasi CPU dan Ambang Batas Operasional
“Max Practical Throughput (Mbps) 520 380 250 2,08x”
Citations
Cites
Wireguard: next generation kernel network tunnel
Donenfeld, J. A. (2017) · Proceedings 2017 Network and Distributed System Security Symposium · doi:10.14722/ndss.2017.23160 · Cited by 243
Baseri, Y., Chouhan, V., & Hafid, A. S. (2024) · Computers & Security · doi:10.1016/j.cose.2024.103883 · Cited by 110
Securing O-RAN Open Interfaces
Groen, J., D’Oro, S., Demir, U., Bonati, L., Villa, D., Polese, M., Melodia, T., & Chowdhury, K. R. (2024) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2024.3393430 · Cited by 72
Akinsanya, M. O., Ekechi, C. C., & Okeke, C. D. (2024) · Engineering Science & Technology Journal · doi:10.51594/estj.v5i4.1076 · Cited by 64
Gentile, A. F., Macrì, D., Rango, F. D., Tropea, M., & Greco, E. (2022) · Future Internet · doi:10.3390/fi14090264 · Cited by 47
Nam, T. S., Hoang, V.-T., & Bui, D.-H. (2023) · IEEE Access · doi:10.1109/access.2023.3306420 · Cited by 35
Full‐mesh VPN performance evaluation for a secure edge‐cloud continuum
Kjorveziroski, V., Bernad, C., Gilly, K., & Filiposka, S. (2024) · Software Practice and Experience · doi:10.1002/spe.3329 · Cited by 31
Anyam, J., Singh, R. R., Larijani, H., & Philip, A. (2025) · Computers · doi:10.3390/computers14080326 · Cited by 30
Comparative Analysis of VPN Protocols at Layer 2 Focusing on Voice Over Internet Protocol
Budiyanto, S., & Gunawan, D. (2023) · IEEE Access · doi:10.1109/access.2023.3286032 · Cited by 27
Orchestrating Isolated Network Slices in 5G Networks
Esmaeily, A., & Kralevska, K. (2024) · Electronics · doi:10.3390/electronics13081548 · Cited by 23
Research works printed but not linked: 3
- DOI: https://doi.org/10.54082/jupin.2313 Vol. 6, No. 1, Februari 2026, Hal. 633-648 p-ISSN: 2808-148X https://jurnal-id.com/index.php/jupin e-ISSN: 2808-1366 DOI: https://doi.org/10.54082/jupin.2313 Vol. 6, No. 1, Februari 2026, Hal. 633-648 p-ISSN: 2808-148X https://jurnal-id.com/index.php/jupin e-ISSN: 2808-1366 DOI: https://doi.org/10.54082/jupin.2313 Vol. 6, No. 1, Februari 2026, Hal. 633-648 p-ISSN: 2808-148X https://jurnal-id.com/index.php/jupin e-ISSN: 2808-1366 studi ini menggeser paradigma benchmarking dari sekadar pengujian teknis menjadi evaluasi kelayakan operasional yang lebih objektif dan terukur. Namun, lanskap performa ini tetap terikat pada karakteristik infrastruktur berbasis AMD Ryzen dengan dukungan instruksi AVX2, yang secara inheren memberikan ruang optimasi lebih luas bagi implementasi berbasis kernel dibandingkan arsitektur tingkat pengguna (user-space). Ketergantungan pada ekosistem Linux Kernel 6.8+ juga menyisakan ruang diskusi mengenai generalisasi hasil pada manajemen memori sistem operasi yang berbeda, dan karenanya skor efisiensi dan model skalabilitas yang dihasilkan harus dipandang sebagai fondasi awal bagi desain jaringan berskala masif. Kedepannya, ketahanan kerangka kerja ini perlu diuji terhadap algoritma kriptografi pasca- kuantum (post-quantum cryptography) yang diprediksi akan menuntut beban kerja jauh lebih berat (Jiang dkk., 2024). Perluasan pengujian pada arsitektur perangkat keras yang lebih beragam, seperti sistem berbasis ARM, akan menjadi langkah krusial untuk memperkokoh validitas model prediktif ini sebagai standar evaluasi global bagi infrastruktur VPN masa depan.
- Hussein, S. G., & Rehman, S. M. F. U. (2025). Protocol efficiency and resource utilization in VPN technologies: A comparative analysis of OpenVPN and WireGuard. Journal of Techniques, 7(4), 37-42. https://doi.org/10.51173/jt.v7i4.2768 · DOI · Paper page
- Tran, N., Hoang, T., & Bui, H. (2023). A hardware architecture of NIST lightweight cryptography applied in IPSec to secure high-throughput low-latency IoT networks. IEEE Access, 11, 89405- 89416. https://doi.org/10.1109/ACCESS.2023.3306420 · DOI
Cited by
Still being checked.
Metadata
- Authors
- Suryaningrat, A., & Rismawan, U.
- Affiliations
- Universitas Dian Nusantara (ID)
- Venue
- Jurnal Penelitian Inovatif
- Publisher
- CV Firmos
- Published
- 2026-02-26
- Volume
- 6
- Issue
- 1
- Pages
- 633-648
- ISSN
- 2808-1366 (ISSN-L)
- 2808-148X
- OpenAlex
- W7132834372
- Cited by
- 0
- References
- 21
- FWCI
- 0.00
- Type
- journal article
- Language
- id
- Open access
- diamond
- doi.org
- Locations
- Jurnal Penelitian Inovatif (publishedVersion, cc-by)
- Topics
- Physical Sciences › Computer Science › Computer Networks and Communications › Cybersecurity and Information Systems
- Physical Sciences › Computer Science › Computer Networks and Communications › Network Security and Intrusion Detection
- Physical Sciences › Computer Science › Hardware and Architecture › Network Packet Processing and Optimization
- Keywords
- WireGuard, IPsec, OpenVPN, throughput, latency, CPU utilization, network security protocols, cloud networking
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.
A language model drew the claims from the paper on 2026-11-21, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-11-21 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
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
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.