A Server Network You Can Measure Before Buying
Most hosting providers ask you to trust a map. We would rather you trust your own measurements. The tables on this page give you a reachable address range inside each data center we use, so the latency numbers you see in your terminal are the numbers your visitors, players or API clients will see once your server is live there.
Our team starts every project with the same question: where are the people who will connect to this machine? A store selling in the Gulf, a game community spread across Eastern Europe and a VPN service with users in Latin America each need a different origin. Once we understand the audience, we help you combine the right location, resources and IP setup so you pay for what the project needs and nothing it does not.
Behind those ranges sits a network built for production work, with Anycast support and enterprise DDoS protection available across our infrastructure. Whether you deploy a small Linux VPS for a landing page or a multi-server stack that spans continents, the same support team helps you choose, test and adjust the setup as the project grows.
The list covers more than 110 VPS locations and a separate set of dedicated server facilities. Some countries have a single data center; others have two, three or four, each with its own upstream connectivity. Testing more than one range in the same country is often the quickest way to find the route that suits your users best.
Reading the Data Center Test IP Table
Each row in the tables names a country and the city where the hardware sits, followed by one or more IPv4 ranges. The last octet, and sometimes the last two, is masked as xxx because those addresses belong to live customer machines. You do not need the exact address to test a route: any host inside the same /24 normally follows the same path through the same upstream, so a test against the published prefix tells you how that facility performs from your side.
Where a country shows several entries separated by a vertical bar, each one is a different data center, labeled DC1, DC2, DC3 and so on. The short letters in brackets are our internal facility codes. They let support identify the exact site when you open a ticket, so quote them if you ask about a range.
A handful of rows show a fully masked address. In those locations we do not publish a test range at the moment, usually because the subnet is shared with a small number of customers or is being changed. Capacity there may still be available; open a ticket and we will tell you what can be tested before you order.
The final column links to the plans for that location, so once a range performs well you can go straight to the matching VPS page and compare cores, RAM, storage and port speeds.
Choosing a VPS Location by Audience Latency
Our aim has always been fast loading, dependable service and the lowest latency we can offer at a fair price. Latency is mostly a matter of distance and routing, so the first decision is not which plan to buy but where your market or your team actually is. Start broad with the continent, narrow it to the region, then use the test ranges on this page to pick the specific country and facility.
Some examples of how that plays out. A WooCommerce store selling to Saudi Arabia and Kuwait gains more from a Riyadh or Dubai origin than from a bigger server in Frankfurt. A Minecraft or CS2 community split between Poland, Czechia and Hungary can sit in any of those capitals and keep most players in a comfortable range. A remote desktop for an accounting team in Karachi feels far more responsive from Islamabad than from Singapore.
If your audience is spread across several regions, one origin will always be a compromise. In that case, put the application where most paying users are, then add a second node or a CDN for the rest. Our multi-location VPS bundles exist for exactly that pattern.
Our standing advice: begin with a virtual private server. It is cheaper to move a VPS than a dedicated machine, so you can confirm a location under real traffic before committing to hardware.
Testing Our IP Ranges From Your Own Network
The most honest latency test is the one you run from the networks your users actually sit on. Ping a published range from your office line, from a mobile connection and, if you can, from a friend or colleague in the target country. A low number from your fiber at home means little if your customers browse on a congested mobile carrier in another region.
Run each test more than once and at different hours. Evening peaks on residential networks can add noticeable delay and jitter that a single midday check will not reveal. Note the average and the worst result, not just the best one, because the slowest responses are what users notice in games, voice calls and remote desktops.
A traceroute against the same address shows the hops your traffic takes. If the path leaves the region and comes back through a distant exchange, you will see it in the hop names and the jump in response time. That is a sign to try another facility in the same country or a neighbouring one.
When you cannot reach testers on the ground, our free ping test from global nodes checks a host from many countries at once, which gives a quick picture of how a range performs well beyond your own desk.
Comparing DC1, DC2 and DC3 Ranges in One Country
Many of our countries have more than one facility. France lists four separate data centers, Brazil, Bulgaria, Greece, Hong Kong, Israel, Japan, Malaysia, Mexico, Pakistan, South Africa, Sweden, Turkey and the UAE list three or more ranges, and countries such as Argentina, Australia, Canada, Egypt, Georgia, Hungary, Iraq, Nigeria, Panama, Saudi Arabia, Ukraine and Uzbekistan offer two. These are not duplicates. Each facility connects to the internet through its own upstream providers, so routes to the same audience can differ by many milliseconds.
That difference matters most for audiences outside the host country. Two Istanbul facilities may perform almost identically for Turkish visitors yet route very differently toward the Gulf or Central Asia. One Tokyo range may reach Southeast Asia more directly than another. The only reliable way to know is to test each range from your target networks and compare.
Facilities also differ in hardware and plan design: storage type, port speed, bandwidth model and available virtualization. When two ranges perform similarly, the plan details on the country page usually decide which one suits the workload.
If you need a presence in two facilities of the same country, for redundancy or to split traffic by carrier, ask support. Spreading servers across separate data centers protects you from a single site's maintenance window.
ISP Carriers and Port Speeds Behind Our DCs
The internet providers that connect our servers to the wider internet were chosen for their reputation and reliable service. Good upstreams are what turn a well-placed data center into a fast one, since they decide how many networks your traffic can reach directly and how often it has to detour through distant exchanges.
Port speed varies by location because bandwidth prices vary enormously around the world. As a rule, each dedicated machine is connected through a 100 Mbps port. In a few data centers the port is lower because transit is very expensive there, while in markets where bandwidth is cheap the connection is faster than 100 Mbps. VPS plans follow the same logic, and each country page states the exact port and bandwidth allowance for every plan.
When you compare locations, look at both numbers: latency tells you how responsive the server will feel, while port speed and monthly traffic tell you how much data it can push. A file mirror or streaming relay needs throughput; a game server or trading bot cares more about consistent low delay.
Our internet speed test is useful for checking your own connection first, so you know whether a slow transfer is the server's limit or your local line.
Clean IP Ranges and Blocklist Checks
Our data centers provide clean IPs and ranges so you can start an online business without landing on a blocklist on day one. A new mail server, a proxy exit point or a payment-facing app depends heavily on the reputation of the address it runs on, and inheriting a range with a poor history can cost weeks of delisting requests.
Before ordering, you can run the published prefix through the common public blocklist lookups and email reputation checkers. If a range you plan to use for outgoing mail or scraping shows old listings, tell us before provisioning and we will look for a cleaner block in that location where one is available.
This page also helps with a different kind of blocking: networks that refuse traffic from certain countries or providers. Some streaming services, banks, game publishers and corporate firewalls filter by IP origin. If your users sit behind such a network, ping or connect to the test range from inside it. A timeout where other hosts answer is a strong hint that the range is filtered there.
Reputation is a shared job. Keeping your server patched, closing open relays and following our terms keeps the range clean for you and for the next customer on the same subnet.
Geolocation of Our Country-Specific IPs
Many customers choose a location precisely because they need an IP that appears in a particular country: local search testing, region-locked content, geo-targeted ad verification, or simply showing local visitors a nearby server. The ranges on this page are allocated to data centers inside the listed countries, which is why we call them geolocated IPs.
Commercial geolocation databases, however, do not all update at the same speed. A range that recently moved between regions or providers can show the correct country in one database and the previous one in another for a while. Services that rely on a single database may therefore see a different location than the one where the hardware actually runs.
Check this before you order if location matters to your project. Our IP lookup tool shows how a range is reported, including country, city and network owner. If the result does not match what the site you care about expects, mention it in a ticket so we can suggest a better range or another facility.
Keep in mind that the city field is often less precise than the country. A Tbilisi or Kyiv address may be reported as the capital or as the registered office of the network owner, which rarely affects country-level targeting.
Requesting a Specific Subnet Before You Order
We keep the listed subnets available, but we cannot guarantee that a new server will be provisioned in exactly the range shown here. Address pools move as customers join and leave, and providers sometimes add new blocks. If your project depends on one particular range, contact support before you place the order rather than after.
Tell us the location, the range you tested and why it matters: a firewall allowlist on a client's side, a geo-check that passed on that block, or a latency result you want to keep. We will confirm whether an address in that range can be reserved for your order or suggest the closest alternative in the same facility.
The same applies when you need several addresses. Proxy operators, agencies with many client sites and mail senders often want additional IPv4 from one subnet or deliberately from different ones. Both are easier to arrange before provisioning than after the server is already running.
If you are latency-sensitive or worried about blocking on a particular network, a short ticket at the start saves a migration later. That is the main reason this page exists: to let you ask the right question while changing your mind is still free.