Archived benchmark data repository. This repository is kept as a historical data sample and benchmark-format reference, not as a live VPS ranking or promotion page.
This repository contains a small VPS benchmark framework and an old BandwagonHost benchmark sample. It is no longer the primary benchmark hub because VPS hardware, routes, stock, pricing and network conditions change frequently.
For current maintained resources, start here:
| Need | Maintained Resource |
|---|---|
| Current VPS benchmark hub and data structure | vps-benchmark-hub |
| BandwagonHost / 搬瓦工 speed test and datacenter notes | bwg-speed-test |
| General VPS tools and server resources | awesome-vps-tools |
| 中文搬瓦工套餐、库存和购买前判断 | BWH Guide |
vps-benchmark-results/
├── benchmarks/
│ └── bwg-cn2-gia-e-2gb.json
└── scripts/
└── run-benchmark.sh
The preserved JSON file is a benchmark-format sample. It should not be treated as a current provider ranking or current performance guarantee.
A public benchmark record should describe the tested environment without exposing private account or server details:
{
"provider": "ExampleProvider",
"plan": "Example Plan",
"region": "Example Region",
"timestamp": "2026-01-01T00:00:00Z",
"tester": "community",
"specs": {
"cpu_cores": 2,
"cpu_model": "Example CPU",
"ram_gb": 2,
"disk_gb": 40,
"disk_type": "SSD",
"os": "Ubuntu 22.04 LTS",
"hostname": "omitted"
},
"results": {
"cpu_score": 2847,
"disk_io_mbps": 485.2,
"network_mbps": 945
}
}Do not publish:
- server IP addresses;
- real hostnames;
- control panel ports;
- account emails;
- API keys, tokens or private credentials;
- affiliate backend screenshots;
- order IDs, invoice IDs or private dashboard screenshots.
The included script is intentionally simple and is kept for historical reference:
bash scripts/run-benchmark.shBefore sharing any generated result publicly, review the JSON output and remove private environment details. The script now writes "hostname": "omitted" by default.
Benchmark numbers are only useful with context:
- CPU results depend on CPU generation, host load and virtualization settings.
- Disk results depend on cache, filesystem and test method.
- Network results depend on route, time of day, remote test endpoint and provider shaping.
- Old data should be used as a format example, not as a purchase decision by itself.
This repository is archived in practice:
- No live ranking is maintained here.
- No direct affiliate or referral links are included here.
- New benchmark methodology and provider comparisons should be added to vps-benchmark-hub.
MIT