content
hide
Preface: Distance and optimization, which is more important?
When choosing a VPS, we often hesitate between a local, standard connection and a more distant, optimized connection. To visually demonstrate the performance differences between the two at different times of day, I conducted a comparative test. This article is the first part, documenting the results from the off-peak hour (8:00 AM).
- VPS data used in the video:
- Japan:https://paste.spiritlhl.net/#/show/tyZy7.txt
- Singapore:https://paste.spiritlhl.net/u/MDyu7W.txt
- US CC:https://paste.spiritlhl.net/#/show/I2RQW.txt
- US Residential IP:https://paste.spiritlhl.net/#/show/RTECZ.txt
- Bricklayer DC1 CN2 GIA+ CMIN2:https://paste.spiritlhl.net/#/show/uEW4q.txt
- If you are willing to support me, you can consider placing an order to purchase VPS through my invitation link~
- Bricklayer:https://dlj.playfulzone.top/zfnw9k
- US Residential IP:https://dlj.playfulzone.top/wvk96m
- Cheap VPS:https://dlj.playfulzone.top/9yp9mz
- SoftBank Japan/iij:https://dlj.playfulzone.top/ghxgwr
Test environment
| project | Configuration |
|---|---|
| Local network | Beijing Mobile 500Mbps (measured speed: 585Mbps / 108Mbps) |
| Node 1 (Japan) | Tokyo, regular line (IIJ), 10Gbps bandwidth |
| Node 2 (Singapore) | Singapore, standard line, 10Gbps bandwidth |
| Node 3 (US-1) | Silicon Valley, regular line (residential IP), 1Gbps bandwidth |
| Node 4 (US-2) | United States, standard line, 1Gbps bandwidth |
| Node 5 (US-Optimized) | Los Angeles, Bandwagon Host (China Telecom CN2 GIA, China Mobile CMIN2), 2.5Gbps bandwidth |
Speed test results during off-peak hours (8 a.m.)
1. Tokyo, Japan (short-distance ordinary line)
- Download speed: 576 Mbps
- analyze: It basically maxed out my home broadband limit and performed very well.
2. Singapore (short-distance ordinary route)
- Download speed: about 580 Mbps
- analyze: It also fully utilized the home broadband limit, which is comparable to the performance of the Japanese node.
3. Silicon Valley, USA (Long-distance Ordinary Line)
- Download speed: about 250 Mbps
- analyze: The speed is about half of the local bandwidth, which is acceptable for ordinary lines during off-peak hours.
4. United States (Long-distance ordinary line)
- Download speed: about 240 Mbps
- analyze: There were certain network fluctuations, but the final speed was basically the same as the previous ordinary US line.
5. Los Angeles, USA (Long-Distance Optimized Route)
- Download speed: 580+ Mbps
- analyze: Despite being the farthest geographical location, it easily maxed out my home broadband limit thanks to the optimized lines.
Off-peak testing summary
| Line Type | Geographical location | Off-peak download speed (based on 585Mbps local bandwidth) | in conclusion |
|---|---|---|---|
| Ordinary lines | Close range (Japanese/Singapore) | ~580 Mbps | Fully utilize local bandwidth |
| Ordinary lines | Long distance (US) | ~250 Mbps | About half of the local bandwidth |
| Optimize routes | Long distance (US) | ~580 Mbps | Fully utilize local bandwidth |
From the test results during off-peak hours, we can draw the following conclusions:
- For 500M-level home broadband,Long-distance optimized lines (such as the US CN2 GIA) and short-distance ordinary lines (such as Japan and Singapore) can both reach the bandwidth limit in terms of download speed, with almost no difference in performance.
- For ordinary long-distance lines, even during non-peak hours when the network is idle, the speed loss is quite obvious and can only reach about half of the theoretical bandwidth.
- From the perspective of latency, close-range nodes are undoubtedly more advantageous.
Next issue preview: The real test will come during the evening rush hour. How will the performance of these routes change then? Will the value of optimizing routes become more apparent? Stay tuned for the second part of the test.
