Istanbul
Assess your connection route for projects focused on Türkiye.
Opening the test connects you to an external testing service.
Starting a measurement may use your internet data allowance.
Istanbul and New York connection test.
Check your connection under real conditions when choosing a location. The tools below open the existing test endpoints; results vary with your connection, route and current load.
New York
Assess your connection route for projects focused on America.
Opening the test connects you to an external testing service.
Starting a measurement may use your internet data allowance.
These tools help evaluate locations; they do not constitute a performance or latency commitment for a particular plan. Verify your order’s data center and connection details at the quote stage.
How should you assess the result?
Look beyond a single measurement.
Latency
Lower latency can improve response times in interactive applications. Measuring from the country where your users are located is more meaningful.
Connection conditions
Wi-Fi, VPNs, mobile networks and background downloads affect results. Compare at different times and, if possible, with a wired connection.
Application performance
A good connection alone does not mean a fast site. Database queries, caching, images and resource usage should also be measured.
GUIDE TO READING MEASUREMENTS
Ping, jitter and speed measure different things.
Record the test’s metrics with their own units. Latency measurements in a browser may not use the same protocol as command-line ICMP ping.
| Metric | What does it tell you? | What does it not tell you? |
|---|---|---|
| Latency / RTT · ms | Shows the time for a request to travel out and its response to return. The user’s network and route affect the result. | Does not measure the page’s database, images or full loading time. |
| Jitter · ms | Shows variation between consecutive latency measurements. Consistent responses matter alongside a low average. | Cannot be inferred from one lowest ping value; calculation methods may differ. |
| Packet loss · % | If supported by the test, shows the proportion of attempts that receive no response. | A firewall limiting test packets does not alone prove that application traffic is also being lost. |
| Download / upload · Mbps | Shows the transfer speed achieved to the test endpoint under specific conditions. | Does not verify the disk speed, CPU power or bandwidth allocated to the server you will purchase. |
Compare locations under the same conditions.
A five-step measurement record
- Start with the same device and network; note VPN, background download and Wi-Fi status.
- Run the Istanbul and New York tests consecutively under the same connection conditions.
- Measure each endpoint several times. Record the typical value and range of variation instead of selecting only the best result.
- Repeat at busy and quiet times. If possible, also sample from another network where your audience is located.
- Make the final decision alongside your application’s actual page or API response; do not treat a network measurement alone as capacity advice.
How would you interpret an example result?
Illustrative example only: Suppose endpoint A measures 25, 26 and 80 ms, and endpoint B measures 35, 36 and 37 ms. A has the shortest minimum response; B is more consistent across these three measurements. This example does not measure our actual locations.
Three attempts do not establish service quality throughout the day. If variation appears only on Wi-Fi, try again on a wired network first. If speed drops at both endpoints, examine your local connection, VPN or concurrent downloads.
Download a blank measurement table · CSV →If the test does not open, try the separate-tab link. An inaccessible external test endpoint does not mean your hosting service is also down. You can include the destination address, time and connection type with the result in your support ticket.
Let’s choose the right location for your project together.
Share your users’ locations, application and resource needs.