Under Ubuntu 26.04 Self-built VPN speed test report: ZgoCloud + Wstunnel + Wireguard solution experience and comparison guide
Preface
In the choice of network acceleration tools, many friends may be looking for a stable and efficient solution. Whether it is a self-built VPN or using a business service, the core goal is to get a reliable Internet experience. Today, I would like to share with you a set of VPN architecture that I built, through the measured data (based on 14:00 Beijing time, Chengdu, Sichuan, China mobile environment) and comparison test methods, combined with9 key screenshots embedded in chronological order, help readers intuitively associate the operation steps and results, and clearly understand the performance of the plan. Whether you are using other self-built programs or business services, you can follow the steps in the text to make your own choices with data.
The measured evening rush hour has been updated, see the third stage and subsequent comprehensive comparison
You may be concerned: is this solution still stable during peak traffic periods? Tonight (June 2) At around 21:00, I will test again in the same environment, compare the data differences between day and night peaks, and verify its pressure resistance.
The secret of the core technology stack
In order to balance stability, anti-interference ability and low latency, I chose the following combination of technology:
- Server nodes: zgocloud los angeles, usa
Route advantages: Adopt the Los Angeles triple network optimization node (CN2 GIA + 9929 + Cmin2), which is optimized for mainland China to alleviate peak congestion. - Transport Protocol: WSTUnnel + Wireguard
Technical principle:WireGuard is responsible for core encryption transmission and tunnel establishment,WSTUnnel encapsulates WireGuard traffic as HTTPS traffic, taking into account concealment and penetration. - Client Management: Clash Verge Rev (Ubuntu)
Functional highlights: intelligent shunting rules, realizing the inductive switching of domestic direct connection and foreign agents.
2. Explanation of the measured environment
The test environment is as follows to ensure that the data can be reproduced:
- Operating System: Ubuntu Linux (26.04)
- Physical location: Chengdu, Sichuan, China
- Broadband service provider: China Mobile (China Mobile), gigabit bandwidth
- Test tool: Speedtest.net (Chrome browser version)
- Test time: around 14:00 Beijing time
Note: If the SpeedTest app is not used due to system restrictions, the Android side data will be supplemented later.
(Figure 1: Screenshot of the test environment configuration)
(Screenshot description: Display Ubuntu system version, network adapter information, label test time and device information.)
Note: If the SpeedTest app is not used due to system restrictions, the Android side data will be supplemented later.as shown in Figure 9
Open SPEEDTEST official website, toggle the language to Chinese (Simplified). as shown in Figure 2
During the test, the browser requires the relevant location permissions, set to allow.
3. Phase 1: Local network benchmark test (control group)
After disconnecting the agent, the three test results are as follows:
- Local test #1
Download: 549.98 Mbps, upload: 95.40 Mbps → Typical performance of mobile broadband (fast downward, uplink limited).
(Figure 3: Local Benchmark #1 Screenshot)
- Local test #2
Download: 533.25 Mbps, upload: 95.70 Mbps.
(Figure 4: Local Benchmark #2 Screenshot)
- Local test #3
Download: 467.04 Mbps, upload: 94.52 Mbps.
(Figure 5: Local Benchmark #3 Screenshot)
📊 Local benchmark summary:
- Average downloads: ~530 Mbps
- Average upload: ~95 Mbps
- Average delay: ~45 ms
4. Stage 2: Self-built VPN measured (experimental group)
After connecting the VPN, three test results:
- VPN Test #1
Analysis: Although it spanned the Pacific Ocean, the download still ran to 80.03 Mbps and uploaded 55.53 Mbps. For the cross-ocean connection, this achievement is very good.
(Fig. 6: Screenshot of VPN speed measurement #1)
- VPN Test #2
Analysis: extremely small fluctuation, download 77.64 Mbps, upload 44.83 Mbps. Idle Latency is controlled at 173ms, which means there is almost no lag in browsing the web.
(Figure 8: VPN delay test screenshot)
- VPN Test #3
Analysis: The third test is still stable, and the upload has even returned to 81.05 Mbps. This demonstrates the excellent performance of the CN2 GIA + 9929 line during the afternoon (non-late night).
(Figure 7: Detailed data screenshot of VPN speed measurement #3)
(Screenshot description: SpeedTest detailed report page, showing download/upload rate graph and complete test results.)
5. The third stage: evening rush hour stress test (21:00 Beijing time supplementary test)
According to the needs of readers, under the same environment (Ubuntu + Chengdu Mobile Gigabit), in Around 21:00 in the evening Retest, compare day (14:00) data, and verify the anti-pressure ability of the scheme.
🕘 Local Benchmark for Evening Peak (Control)
Disconnect the VPN and measure the speed three times in a row:
Local evening rush hour #1
Download:450.11Mbps, upload:95.76Mbps, ping: 46ms
→ The downside of the evening peak has dropped from 530Mbps during the day to 450Mbps, a drop of about 15%.
(Figure 10: Screenshot of the local speed measurement of the evening peak)
Local evening rush hour #2
Download:427.54Mbps, upload:94.60Mbps, ping: 33ms
(Figure 11: Screenshot of local speed measurement in the evening peak)
Local evening rush hour #3
Download:447.27Mbps, upload:95.92Mbps, ping: 34ms
(Figure 12: Screenshot of local speed measurement at night peak)
📊 Summary of local benchmarks for evening rush hour(contrast with day):
| Time period | average download | average upload | Average ping |
|---|---|---|---|
| Daytime (14:00) | ~530 Mbps | ~95 Mbps | ~45 ms |
| Evening rush hour (21:00) | ~441.6 Mbps | ~95.4 Mbps | ~37.7 ms |
Analysis: The local download speed of the evening rush hour has decreased slightly (about 16%), but the upload speed is almost unchanged, and the ping is even slightly improved. It shows that the overall load of Chengdu Mobile Broadband during the evening rush hour is acceptable, and there is no serious congestion.
🌙 Evening Peak VPN Measured (Experimental Group)
After connecting the VPN, the speed measurement is three times in the same environment:
VPN Evening Peak #1
Download:62.17Mbps, upload:54.29Mbps, idle delay:185ms
→ It is about 20% lower than that of the daytime VPN download (72 Mbps), but it is still much higher than the 4K streaming demand.
(Figure 13: Screenshot of the evening peak VPN speed measurement #1)
VPN Evening Peak #2
Download:54.58Mbps, upload:73.46Mbps, idle delay:187ms
(Fig. 14: Screenshot of the evening peak VPN speed measurement #2)
VPN Evening Peak #3
Download:55.60Mbps, upload:78.68Mbps, idle delay:197ms
(Figure 15: Screenshot of the evening peak VPN speed measurement #3)
📊 Evening Peak VPN Summary(Compared to a daytime VPN):
| Time period | Average Downloads (VPN) | Average Upload (VPN) | average idle delay |
|---|---|---|---|
| Daytime (14:00) | ~72.3 Mbps | ~60.7 Mbps | ~173 ms |
| Evening rush hour (21:00) | ~57.5 Mbps | ~68.8 Mbps | ~189.7 ms |
observe: Evening peak VPN upload speed (average 68.8 Mbps) is higher than daytime (60.7 Mbps). It may be the fluctuation of SpeedTest speed nodes, or the uplink resources are more abundant at night peaks. In any case, this upload speed has far exceeded video conferencing needs.
6. Core conclusion: Daytime VS evening peak, how is the plan performance?
📊 Comprehensive performance comparison table
| Scenes | local download | vpn download | upload locally | VPN upload | Idle Delay (VPN) |
|---|---|---|---|---|---|
| Daytime (14:00) | 530 Mbps | 72.3 Mbps | 95 Mbps | 60.7 Mbps | 173 ms |
| Evening rush hour (21:00) | 441.6 Mbps | 57.5 Mbps | 95.4 Mbps | 68.8 Mbps | 189.7 ms |
✅ Core conclusions (three articles)
- absolute performance: Evening rush hour VPN download is stable at 55–62 Mbps, upload to 55–78 Mbps. This bandwidth can easily support 4K video streaming, large file downloads, and multi-channel high-definition video conferencing. It is about 20% lower than the day download, but it is still more than enough.
- Relative loss rate is reasonable: The daily VPN loss is about 86.4% (530→72.3), and the evening peak loss is about 87.0% (441.6→57.5).The tunnel efficiency is almost identical, the download drop of the evening peak is mainly derived from the speed attenuation of the local basic network (530→441.6).
- Delay performance is stable: The evening rush hour idle delay average is 190 ms, which is only about 10% higher than the day (173 ms). There is almost no perceived difference in web browsing, and only competitive games (US servers) that are extremely sensitive to delays may feel a slight increase in operation lag.
Interpretation of the key parameters of self-built VPN measured (comparison of data during the day and evening peak)
In addition to the up and down speed, the three delay parameters in the screenshot are very important. Combination below Day and evening rush hour The actual data to illustrate:
| Parameter | Meaning | Daytime VPN | evening rush hour vpn | Evaluate |
|---|---|---|---|---|
| Idle Latency | ping value, reaction network reaction time | 173 ms | 185–197 MS | All excellent. Below 200ms is the first cross-ocean first echelon. |
| Download latency | Delay jitter when downloading | 186–191 ms | 193–199 MS | normal. The evening rush hour rises slightly, which does not affect daily use. |
| Upload Latency | Delay jitter when uploading | 333–481 MS | 216–349 MS | The evening peak upload jitter is better than the day (up to 349ms in Figure 13–15), indicating that the line quality is stable. |
If you need to have a video conference at full speed uploading files, it is recommended to limit the upload bandwidth on the router or computer (such as a limit of 30 Mbps) to avoid the stuck feeling caused by high upload delays.
8. Why are there two results of ‘Suzhou Mobile’ and ‘Los Angeles’?
Whether it is daytime or evening rush hour, the principle of local speed measurement and VPN speed measurement point to different nodes is consistent:
- Figure 10–12 (local): SpeedTest automatically matches to
jsqy (suzhou)node. This is the backbone network node of China Mobile in Jiangsu Province. The physical distance is far but it belongs to the high-speed channel of the operator, so it has run at a speed of 400+ Mbps. - Figure 13–15 (VPN): SpeedTest intelligently recognizes my current export IP in the United States, so it is automatically matched to
Netlab Global / ReliableSite Hosting (Los Angeles)node.
Conclusion: This is the real transnational speed! If you can connect to the Suzhou node to measure 400+ Mbps when the VPN is turned on, then it is a fake data (because the traffic cannot be instantly detoured back to China before going out). 55–62 Mbps (evening peak) is the real overseas bandwidth that I can actually use.
9. Differences in the experience difference between different operators during peak hours (reference of telecom/unicom users)
- Mobile broadband features: The local download of the evening peak has dropped from 530 to 442 Mbps, and the drop is controllable. VPNs will eventually have a bandwidth of about 55–60 Mbps, and the upload performance is even better than that during the day. The data in this paper has a high reference value for mobile users.
- Potential Advantages of Telecom/Unicom: Since the CN2 GIA / 9929 line has better routing optimization for telecom and China Unicom,Line congestion control may be better during evening rush hour, the actual available bandwidth may be higher than the test data in this mobile environment.
Summary: If you use Telecom or China Unicom,The evening peak experience may be better than mobile data in this article; If it is the same mobile user, the data in this article has direct reference significance.
10. Granting people to fish: a comparison test guide
To verify the performance of any VPN or proxy service during peak hours, refer to the following steps:
- Fixed speed measurement node
Avoid automatic selection of servers, manually specify the same node (such as using Suzhou jsqy during the day, and using Los Angeles Netlab or the same third-party node after VPN). - Three consecutive measurements
A single result was easily affected by fluctuation, and the average was measured 3 times in a row. - Focus on core indicators(Cross-Ocean Excellent Standard):
| Parameter | Meaning | Cross-Ocean Excellent Criteria |
|---|---|---|
| download | download bandwidth | > 50 Mbps (daytime > 60) |
| upload | Upload bandwidth | > 20 Mbps |
| Idle Latency | Idle Delay (Ping) | < 200ms |
- Be sure to measure the local benchmark at the same time
Only by knowing the real speed of the local network at the same time period can we judge whether the loss of the VPN is reasonable.
11. Summary
The measured data shows (including the evening rush hour 21:00 supplementary test),ZGoCloud + Wstunnel + Wireguard On the premise of sacrificing some local bandwidth,Delivering 55–60 Mbps transnational download bandwidth remains stable during peak hours, the upload is 55–78 Mbps, and the delay is controlled within 190 ms.
- Available in the daytime of about 72 Mbps and about 57 Mbps in the evening peak, both far exceed the needs of 4K streaming (25 Mbps) and video conferencing (10 Mbps).
- The tunnel efficiency does not deteriorate with the time period, and the speed of the evening peak decline mainly comes from the attenuation of the basic network speed of the local operator (530→442 Mbps).
- It is suitable for users who pursue a stable cross-border network experience, especially mobile broadband users with high reference value; Telecom/Unicom users are expected to achieve better performance.
If you want to build or verify other solutions by yourself, the test method in the article can help you make your own judgment with data.
Preview: The evening peak test (including the complete configuration tutorial) on the Android mobile terminal will be written separately, and it is expected to be released within this week. Welcome to pay attention.
Stop the Hassle | Custom WireGuard VPN Setup Service
I have been testing various network optimization solutions for a long time. My personal setup has been running continuously for over a month with absolutely zero downtime. If you are tired of troubleshooting and dealing with complex server configurations, feel free to reach out for a custom solution.
What I Offer:
✅ Remote Setup Service: I will deploy an exclusive VPN on your own server. You keep full control of your data—set it once and forget about it.
✅ Free Trial: New users can apply for a one-month free trial of my self-hosted node, so you can personally experience the speed and stability.
✅ Performance Guaranteed: Deeply optimized routing rules to completely eliminate lag and connection timeouts in daily use.
Contact Me:
Telegram: @shuijingwan
WeChat: 13980074657
Email: shuijingwanwq@gmail.com
