The physical distance between your server and your visitors sets a floor on how fast every page can respond. Picking the right data centre is one of the cheapest speed wins there is, and it is hard to change later, so it is worth measuring before you order.
Before you start
- Know where most of your visitors are. Your analytics tool shows visitors by country.
- A terminal for
pingandmtr(see the network diagnostics toolkit). - MdmHost locations: Germany, USA and Singapore.
Why location matters
Light in fibre covers about 200 km per millisecond, and real routes are longer than straight lines. A rough guide to round-trip latency:
| From → to | Typical round trip |
|---|---|
| Same region (e.g. Paris → Germany) | 5–25 ms |
| US East → US West | 60–75 ms |
| Europe → US East | 80–100 ms |
| Europe → Singapore | 150–200 ms |
| US → Singapore | 170–230 ms |
A new HTTPS connection needs several round trips (TCP, TLS, then the request), so 150 ms of extra latency can add half a second before the first byte arrives. Background in what is latency.
Measure from your visitors' side
- Ask our support team for a test IP in each location, or use the location test addresses listed on our site if available.
- From a computer in your main market, run 50 pings to each:
ping -c 50 TEST-IP- Note the average and the packet loss. Loss above 1% at the destination is a warning sign.
- Run an MTR to the best one or two candidates and read it with how to read an MTR report.
- If your audience is spread out, ask friends or customers in other countries to run the same test, or use an online multi-location ping tool.
How to decide
- Mostly Europe, Middle East or Africa: Germany.
- Mostly North or South America: USA.
- Mostly South and Southeast Asia or Australia: Singapore.
- Global audience: host where your biggest market or your admin team is, then add a CDN for static files; see set up Cloudflare CDN with cPanel.
- Legal or data residency rules (for example GDPR expectations from EU clients) can outweigh a few milliseconds.
Tip: For SSH, RDP, game servers and APIs that are called many times per page, latency matters even more than for a cached website. Put those servers as close to the users as possible.
Troubleshooting
- The nearer location pings slower: routing is not always geographic. Your ISP may hand traffic to a carrier that takes a detour. Test at different times of day and send MTRs to support.
- Latency is fine but the site is slow: the bottleneck is the application, not the network. Compare
curlconnect and TTFB times, then see the WordPress speed guide.