NVMe · KVM · Unmetered
Cloud Servers in 36 Locations
Put a server close to the people who use it. Our global cloud range runs KVM virtual machines on NVMe storage in 36 worldwide locations, each with root access, a dedicated IPv4 address and unmetered bandwidth, all ordered, billed and supported from Pakistan.

The global range at a glance
Four facts that define our cloud servers
Everything else on this page builds on these. If any of them is a dealbreaker for your project, our Karachi and Islamabad VPS ranges or a dedicated server may be the better route.
Why a global range
Cloud servers for projects that live outside one country
A server in Pakistan is the natural home for a Pakistani audience. But many of the people we work with build for somebody else: a freelancer whose clients are in Europe, a software house shipping an app used in the Gulf and North America, a trader whose platform needs to sit near a broker, or a game community scattered across several continents. For them, the location of the server matters more than its country of purchase.
Our global range exists for that situation. You choose a location from the list of 36 on the order page, and the server is deployed there with the same basic shape everywhere: KVM virtualisation, NVMe disks, a dedicated IPv4 address and bandwidth that is not counted against a monthly cap. Billing, support and the client area stay the same no matter where the machine runs.
The word cloud is used loosely across the industry. Here it means virtual servers you can place in many regions, sized by plan and billed monthly in US dollars. It does not mean a managed platform that scales itself; you still run the operating system, the firewall and the software. That honesty matters when you plan who will look after the machine at two in the morning.
Global cloud plans
NVMe cloud server configurations and monthly pricing
Compare vCPU, memory and NVMe storage across the range. Prices are monthly in US dollars and update automatically from our main catalogue. The location list appears on the order page when you choose a plan.
Global DCs
| Core(s) | RAM | Storage | Bandwidth | OS | Price | Order |
|---|---|---|---|---|---|---|
| 2 | 2 GB | 40 GB NVMe | Unlimited | Linux | $17.95/mo | Order 2 2 GB 40 GB NVMe |
| 2 | 4 GB | 60 GB NVMe | Unlimited | Linux / Windows | $25.95/mo | Order 2 4 GB 60 GB NVMe |
| 4 | 6 GB | 70 GB NVMe | Unlimited | Linux / Windows | $39.95/mo | Order 4 6 GB 70 GB NVMe |
| 4 | 8 GB | 90 GB NVMe | Unlimited | Linux / Windows | $45.95/mo | Order 4 8 GB 90 GB NVMe |
| 6 | 10 GB | 100 GB NVMe | Unlimited | Linux / Windows | $66.95/mo | Order 6 10 GB 100 GB NVMe |
| 8 | 12 GB | 150 GB NVMe | Unlimited | Linux / Windows | $79.95/mo | Order 8 12 GB 150 GB NVMe |
| 10 | 14 GB | 170 GB NVMe | Unlimited | Linux / Windows | $99.95/mo | Order 10 14 GB 170 GB NVMe |
| 16 | 16 GB | 210 GB NVMe | Unlimited | Linux / Windows | $129.95/mo | Order 16 16 GB 210 GB NVMe |
| 20 | 24 GB | 310 GB NVMe | Unlimited | Linux / Windows | $199.95/mo | Order 20 24 GB 310 GB NVMe |
International Tier III+ data centers — choose your location from 36 worldwide destinations for the lowest latency to your visitors.
Picking among 36 locations by measuring, not guessing
Geography is a starting point, but network paths do not follow maps. Two cities that look close can be connected through a distant exchange, while a further city may sit on a direct route. The reliable approach is to measure from where your users actually are, using the tools already on your laptop, before you commit a production workload to a location.
From a terminal, mtr or traceroute shows every hop between you and a test address, and ping gives a quick round-trip figure. Ask a few users or colleagues in your target countries to run the same commands and share the output. Patterns appear quickly: one location will usually be consistently better for the region that matters most to you.
If your audience is spread out, choose for the majority and handle the rest with caching. A content delivery network in front of static assets reduces the distance problem for images, scripts and downloads, while the server location decides how quickly dynamic pages and API calls respond. Get the dynamic part right first, because that is the part caching cannot fix.
Keep legal and contractual requirements in view as well. Some clients specify that their data must stay in a particular jurisdiction, and some industries have their own rules. When that applies, the decision is made for you, and the measurements simply confirm which of the permitted locations performs best. Our team can share the current location list on request.
What NVMe storage changes for databases and builds
NVMe drives connect over PCI Express rather than the older SATA interface, which allows far more operations to be in flight at once. In practice, the benefit is not only raw speed but low latency under many small random reads and writes, which is exactly the pattern databases, package managers and container image layers produce.
You notice it in ordinary tasks. A composer install or npm ci completes faster because thousands of small files are written quickly. Database queries that miss the cache return sooner. Restoring a backup or rebuilding a search index takes minutes rather than a long coffee break. These gains compound when several of them happen on the same server at once.
Fast storage does not replace memory. If your database working set does not fit in RAM, NVMe makes the shortfall less painful, but a larger plan with more memory is still the proper fix. Watch the cache hit ratio in your database and the available figure in free -h; when either looks stressed during normal traffic, upgrade the plan rather than relying on the disk.
If you want to see the difference for yourself, a quick fio test with a random read and write profile gives comparable numbers between servers. Run it on a fresh machine before deploying anything, keep the results, and repeat the test later if you ever suspect a storage problem. Baseline numbers turn a vague feeling into something our team can investigate.
Designing a service that spans more than one region
Once servers can live anywhere, it becomes tempting to place one in every region. Resist that until you have a reason, because every extra server is another machine to patch, monitor and back up. A sensible first step is two servers: a primary in the location nearest your main audience and a second elsewhere, used for backups and as a standby.
For read-heavy sites, the second server can serve traffic too. A database replica in another region answers read queries locally, while writes still go to the primary. Static files can be synchronised with rsync on a schedule. This keeps the design simple: there is still one source of truth, and failure of the second server never loses data.
Failover needs DNS planning. Set a modest TTL on the records that point to your primary, perhaps five minutes, so that a change propagates quickly when you need it. Check what resolvers see with dig +short yourdomain.com from different networks. If you use the DNS management included with our domains, you can update these records from the client area.
Write down how you would switch to the standby, then rehearse it once. The steps are usually short: promote the replica, update DNS, confirm that the application can write. The rehearsal reveals the forgotten details, such as a cron job that only exists on the primary or a firewall rule that blocks the new database role, before a real outage forces the discovery.
Making good use of unmetered bandwidth without waste
Unmetered bandwidth removes the monthly transfer bill, which changes how you can design some systems. Off-site backups can run every night instead of every week. Large downloads, container images and software updates can be served directly from your server. Media-heavy sites no longer need to squeeze every file to stay under an allowance.
It does not remove the port limit. Every server still has a maximum rate at which data can leave it, so a sudden burst of downloads is shared among the users requesting them. If you expect large simultaneous demand, combine the server with a content delivery network for static files, and keep the server for work only it can do.
Efficient transfer still helps your users, even when it costs you nothing extra. Enable compression for text responses, set sensible cache headers, and serve modern image formats. Every byte that does not travel is a byte that does not wait on a slow mobile connection at the other end, and that is where most perceived slowness comes from.
Watch for unexpected traffic. A sudden rise in outbound transfer can be a sign that the server has been compromised and is being used to send spam or attack others. Tools such as vnstat or iftop show what is moving in and out. Investigate unusual patterns quickly; abuse that originates from your server is your responsibility, unmetered or not.
Turning a fresh cloud server into a working host
The welcome email gives you the IP address and the initial credentials. Log in over SSH, change the password, add a personal user with sudo rights and install your public key. Then edit the SSH daemon configuration to refuse password logins, and restart it from an existing session so that a typo cannot lock you out completely.
Install updates and a firewall next. On Ubuntu, ufw allow OpenSSH followed by ufw enable is a quick start; on AlmaLinux or Rocky Linux, firewalld is already present. Add only the ports you need, such as 80 and 443 for a web server. Fail2ban or a similar tool adds automatic blocking of addresses that keep failing to log in.
Decide how the server will be rebuilt if it ever has to be. A short shell script, an Ansible playbook or a cloud-init file that installs your packages and copies your configuration is worth writing on day one. With 36 locations available, a repeatable build means a second server in another region is an afternoon of work rather than a week.
Last, connect monitoring and backups before any real traffic arrives. An external uptime check tells you when the server stops answering; a backup job sending data to a server in a different location protects you from mistakes and hardware faults. With those in place, you can deploy your application knowing that problems will be noticed and recoverable.
Same company, different tools
Cloud servers next to our Pakistan VPS and dedicated machines
All three are servers with root access. The difference is where they run and what you are paying for. Use the table to narrow the choice, then read the matching page for detail.
| What matters to you | Global cloud range | Karachi or Islamabad VPS | Dedicated server |
|---|---|---|---|
| Server location | 36 worldwide locations | Inside Pakistan | Pakistan or the Netherlands |
| Storage | NVMe | See the plan table on the VPS page | Depends on the configuration |
| Monthly transfer | Unmetered bandwidth | Shown in the VPS plan table | 50 TB in the Netherlands; shown per plan in Pakistan |
| Hardware | Virtual machine on shared host | Virtual machine on shared host | Whole physical machine |
| Typical fit | International audiences, multi-region setups | Pakistani users, local data | Heavy, steady workloads |
Read more about VPS hosting in Pakistan, serving Middle East audiences or running trading platforms on a VPS.
Who runs cloud servers with us
Typical projects on the global cloud range
Freelance developers often keep a small cloud server near each major client: a staging copy of a site in the client region, so the client sees realistic speed during reviews, and the production deployment once the work is approved. Because billing sits in one client area, keeping track of several servers across regions stays manageable.
Software teams build in Pakistan and deploy wherever their users are. A Laravel or Django application, a Node API behind Nginx, or a set of Docker containers moves easily between our locations, because the virtual hardware and the operating system choices are the same everywhere. Continuous deployment from a Git host works the same way regardless of where the target sits.
Network and infrastructure people use the range for tools that benefit from vantage points: monitoring probes in several regions, DNS servers spread across continents, or VPN endpoints for remote staff. Others simply want an off-site backup target outside Pakistan. For each of these, a modest plan in the right place is worth more than a large plan in the wrong one.
Cloud FAQ
Cloud server questions from developers and sysadmins
Can I move a cloud server from one location to another after ordering?
Each server is deployed in the location you pick at checkout, and its IP address belongs to that site. Moving usually means ordering a server in the new location, copying your data across with rsync or a backup restore, switching DNS, and then cancelling the old one. Open a ticket before you start and the team will confirm the cleanest way for your particular plan.
Is Windows available on the cloud range, or only Linux?
Linux is available on every plan, and Windows is offered on most of them. Smaller configurations may be Linux only, because Windows Server needs more memory to run comfortably. The order page shows which operating systems can be selected for each plan; if the one you need is missing, ask our team before paying so that we can suggest the nearest plan that supports it.
What does unmetered bandwidth mean in day-to-day use?
It means your monthly data transfer is not counted against a fixed allowance, so you do not receive an overage bill after a busy month. Your port speed still sets how quickly data can move at any moment, and the acceptable use rules still apply. For normal websites, APIs, game servers and backups, unmetered transfer simply removes one number you would otherwise have to watch.
How are cloud server invoices billed and paid from Pakistan?
Cloud servers are billed monthly in US dollars through the client area. You can pay by bank transfer to Standard Chartered, Bank Alfalah or HBL, by Easypaisa, JazzCash, UBL Omni, Mobicash or U Paisa, by Visa or Mastercard, by PayPal, by Western Union or MoneyGram, or with a cheque deposited into our bank account. Cash payments are not accepted.
Do I get a dedicated IPv4 address on each cloud server?
Yes. Every server in the range comes with its own dedicated IPv4 address, so you can point DNS records straight at it, run your own mail server if you understand the reputation work involved, and install a certificate for your domains. If a project needs extra addresses, describe what they are for in a ticket and our team will tell you what is possible.
Can I upgrade a cloud server later without rebuilding it?
Upgrades to a larger plan are usually possible within the same location, and in many cases the disk and data carry over. The exact process depends on the plan, so open a ticket with your current plan and the one you want. Take a fresh backup first anyway: any change to storage size is a good moment to confirm that your restore procedure works.
Will you manage the software on my cloud server for me?
Cloud servers are self-managed: you have root access and choose every package on the machine. Our team helps with network issues, the virtual hardware and questions about your plan, and we are glad to point you in the right direction on configuration. If you want a fully managed website instead, shared or WordPress hosting on our PKR plans may suit you better.
Is there a refund if I cancel a cloud server in the first week?
No. Once a server has been provisioned we do not issue refunds, whatever its age. To avoid surprises, start with a smaller plan in the location you are considering, run your own tests, and scale up when the results are what you expect. The team can also answer questions about locations and operating systems before you place the first order.
PKDomain.net · help in English & Urdu
Choose a location and launch a cloud server
Pick a plan, select one of the 36 locations on the order page and pay by any accepted method. If you are unsure which location suits your audience, open a ticket and describe where your users are.