Poland, Germany, France or the UK: choosing a country for your server

This article is for an administrator or manager who is ordering a dedicated server and has to decide which of the four countries to place it in. By the end you will know where to put the primary server, where to put the standby and how to check the response time from your own network.

The short answer#

• The primary server goes in Poland. Of the four countries, it gives the lowest response time for staff and customers in Ukraine.
• The standby goes in Germany or France. Another country means a separate data centre, separate power and separate network links.
• The United Kingdom is for services aimed at British users, and serves as one more independent site.

Below: what this advice rests on, when it makes sense to depart from it, and how to check everything on your own network.

What you will need#

• A list of what will run on the server: the accounting system, remote desktops, files, a website, backups.
• An idea of where your users are: an office in Ukraine, staff working from home, customers in Europe or Britain.
• A computer on the network people will work from, to measure the response time. The Windows command prompt or a Linux or macOS terminal is enough.
• The server catalogue: it shows which configurations are in stock in each country. Availability is updated every five minutes.

Response time from Ukraine to each country#

Response time is how long a packet takes to travel from your computer to the server and back. It is measured in milliseconds (ms) with the ping command.

The figures come from our measurements from Ukrainian providers’ networks, October 2026. Your value depends on your provider and route; on mobile internet the response time is higher. The 15–25 ms to Poland is roughly half the response time to Frankfurt or Amsterdam.

What these milliseconds mean in daily work#

• Remote desktop (RDP). Every keystroke and every mouse movement travels to the server, and the picture comes back. At 15–25 ms, working feels almost the same as on a local computer. 40–50 ms is a workable figure too: for its own remote desktops Microsoft names 150 ms as the limit below which the response time should not get in the way of work that involves no video or graphics. But the margin is smaller: once Wi-Fi, mobile internet or a busy link is added, the lag while typing and scrolling becomes noticeable sooner.
• An accounting system whose client runs in the office while the database sits on the server. One user action is dozens or even hundreds of consecutive requests to the server, and the delays add up. A hundred requests at 15 ms each is a second and a half of waiting on the network alone; at 52 ms each it is more than five seconds. That is why such a system is placed as close to the users as possible, or the client is run on the server itself and people work through a remote desktop.
• A website, mail, a file archive, backups, overnight replication. A few dozen milliseconds change almost nothing here, so the country can be chosen for other reasons.

Why the other countries take longer#

Traffic from Ukraine to Germany, France and the United Kingdom first reaches Poland and only then travels on to the data centre it is bound for. So the response time to these countries is the time to Poland plus the path further west. No setting will shorten it: the signal in optical fibre has to cover the distance there and back, every 100 km of the path adds roughly 1 ms to the response time, and cables do not run in a straight line.

Two practical conclusions follow:

• For users in Ukraine, Poland is the closest of the four countries, so that is where the primary server belongs.
• For a standby server, 35–40 ms is not a drawback. In normal times it receives copies and replicas, and people work on it only during an outage at the main site.

How to choose a country: six steps#

1. Find out where your users are. If they are in Ukraine, put the primary server in Poland. If the service is aimed at British users, put it in the United Kingdom. For visitors from Western Europe, Germany or France will do.
2. Split the workload into what is sensitive to delay and what is not. Place remote desktops, the accounting system and telephony closer to people. Copies, archives and database replicas go wherever suits your standby plan.
3. Decide whether you need a standby site. If a long outage is unacceptable for the company, keep a second server in another country: Germany or France. The response times to the two differ by a few milliseconds, so choose the one that has the configuration you need. The options are described in “Standby site: three schemes and what to check”.
4. Check availability in the catalogue. The set of configurations differs between countries: a model that is available in Poland may be missing in France, and the other way round. The catalogue shows what is in stock and the delivery time for each configuration: within 12 or within 72 hours of payment.
5. If the servers have to work together, take the private network into account. On the main lines, servers in different countries can be joined into an isolated network, for example a primary in Poland and a standby in Germany. The Lite and Turbo lines have no private network; whether a particular model has it is stated on the server card.
6. Measure the response time from your own network. The next section explains how.

If your choice of country is affected by requirements on where data must be stored, clarify them with your lawyer before ordering: this article does not cover legal questions.

How to check it yourself#

Our figures are a guide. Your result depends on your provider, so before ordering it is worth measuring the response time from the network people will work on. Ask support for an address to measure against in the country you need; if you already have a server, measure against its address. The examples use the documentation address 203.0.113.10; replace it with the real one.

1. Connect the computer to the network by cable and switch the VPN off for the measurement: Wi-Fi and a VPN add delay of their own.

2. On Windows, open the command prompt (Win+R, cmd, Enter) and run:

   ping -n 20 203.0.113.10
   tracert -d 203.0.113.10

3. On Linux, run in a terminal (on macOS ping works the same way). On many systems mtr has to be installed first; the article linked below explains how, and until then the first command is enough.

   ping -c 20 203.0.113.10
   mtr -rw -c 100 203.0.113.10

   With these options mtr prints nothing for about two minutes: it sends 100 packets and prints the report at the end.

4. In the ping summary, look at packet loss (it should be 0%) and at the minimum and average time.
5. In the output of tracert and mtr the last hop in the list is the server, and the rest are the hops on the way to it. They show on which stretch the time grows: in your network, at your provider or further on.
6. Repeat the measurement in the morning, at the busiest time of the working day and in the evening, and also from the other networks people work from: home, mobile internet.

How to read the output, why loss at an intermediate hop does not yet mean a problem, and how to install mtr are covered in detail in “How to measure response time and the route to a server: ping, mtr, tracert”.

How to check the result#

• Before ordering. Compare your figures with the table. A difference of a few milliseconds is normal. If the response time is twice the figure in the table, or there is packet loss, look at the trace. If the time is already high at the first hop (that is your router), the cause is in your own network: Wi-Fi, the router, a VPN. If it grows further on and stays high all the way to the last hop, show the mtr or tracert output to your provider.
• After the server is delivered. Repeat ping against the server’s address from every network people will work from. Then work for 10–15 minutes the usual way: open a remote desktop, a typical document in the accounting system, a large file, and judge whether it is comfortable to work.
• If there are two servers. Measure the response time between them: it determines which replication scheme will suit you. Support will send the parameters for connecting to the private network.

Common mistakes#

• Choosing a country out of habit or by its name. “A server in Germany” sounds solid, but for an office in Kyiv it means about 35 ms instead of 15. Redundant power and network and network-level DDoS protection are there in all four countries, so the main difference for the user is the response time.
• Putting the primary server for Ukrainian users in the United Kingdom because the configuration you need happens to be in stock there. From Kyiv that is about 52 ms instead of 15, so every request to the server takes more than three times as long: in an accounting system whose client runs in the office, the difference shows at once. It is better to pick a similar configuration in Poland.
• Keeping the primary and the standby server in the same country. A standby site is there for an outage in the data centre or on the links to it, so another country is the more reliable choice.
• Treating a standby site as a replacement for backups. A second server saves you from an outage on the first, but replication will faithfully carry a document deleted by mistake or an encrypted database over to it as well. Backups are needed separately: see the article on the 3-2-1 rule.
• Measuring once and from one place. A measurement over Wi-Fi, a VPN or mobile internet gives a higher figure than a wired desk will see, and a single measurement at night gives a lower one than the busiest time of the working day.
• Taking the figures in the table as a promise. They are the result of measurements from particular networks at a particular time. Providers’ routes change, so check the response time again after changing provider, and simply once every few months.
• Remembering the private network only after ordering. If servers in different countries have to exchange data over an isolated network, choose both from models that have it.

What next#

• Look at the locations page: it shows the response time from Ukrainian cities to Poland and the availability of servers in each country.
• Open the catalogue for the country you need: Poland, Germany, France, United Kingdom.
• Choose a line and a configuration: “How to choose a server line and configuration”.
• Plan the second site with “Standby site: three schemes and what to check”, and the network between the servers with “Private network between servers and 802.1Q VLAN”.
• If you are moving from a Ukrainian data centre, you will need the migration checklist; how ordering works is described in “From request to access”.

Not sure which country suits you? Write to info@united.net.ua or call +38 044 206 08 08. Tell us where your users work and what will run on the server, and we will advise.