Virtual Servers
Launch Linux and Windows servers in minutes, and scale them as you grow
No server is tied to a single machine: if a physical host fails, yours restarts on another one automatically. Your data stays in Türkiye, and your usage shows up on the monthly invoice line by line.
Why now
- It’s 9:00 on Monday morning and the on-premises server won’t boot. The motherboard has to be replaced, the part is three days out, and finance can’t send a single invoice until then. On a virtual server, the same failure is handled by restarting it on another host.
- Campaign day arrives, traffic jumps fivefold and pages crawl. New hardware takes weeks; the campaign lasts two days. On a virtual server you add CPU and memory from the portal, and a second server joins the load balancer in minutes.
- Every server goes out of warranty in the same year, and the replacement quote calls for a large upfront investment. With virtual servers, hardware refreshes and repairs are Nesnet’s job — you pay only for the CPU, memory and disk you use.
- Developers file a ticket for every test box and wait two days. With a ready template and a first-boot script, a server comes up with its users and SSH key already in place, and the team gets straight to work.
- The renewal invoice for your virtualization licence just came in at several times last year’s price. Nesnet builds and runs the virtualization layer itself; your bill shows resources you used, not a licence negotiation.
What changes
| Today | With Nesnet |
|---|---|
| A hardware failure stops work until the spare part arrives. | Your server restarts on another physical host on its own, and because its disk is spread across several hosts, your data is intact. |
| A new server means weeks of purchasing, racking and cabling. | Choose the template, CPU, memory and disk in the portal, and a Linux server is ready to connect within a few minutes. |
| Warranties, virtualization licences and maintenance contracts each renew on their own schedule. | A single monthly invoice lists vCPU, memory, storage, IP addresses and licences as separate lines. |
| One wrong firewall rule locks you out, and someone has to drive to the data center. | The browser console doesn’t depend on the network: open it, fix the rule, carry on. |
| Support means an English-only ticket in an overseas queue. | Nesnet’s support team helps you in Turkish or English. |
From today to live: how it works
- 1
Discovery call
Work through your needs with Nesnet’s team: how many servers, which operating systems, which licences, how much downtime is acceptable. You leave the call with a sizing proposal and a clear view of what drives the price.
- 2
Your organization and network are prepared
Your portal account, resource allocation, gateway and first network are built by Nesnet’s automation in repeatable steps. The gateway firewall starts from deny-all.
- 3
Pilot server
Launch your first server from a template, install your application and test it with your team. If you’re moving an existing server, a non-critical system goes first and is verified.
- 4
Go-live
Allocate a public IP and publish your service with a firewall and NAT rule. Migrated systems switch over at the agreed time, with the rollback step written down in advance.
- 5
Continuous monitoring and growth
The platform’s hardware and layers are monitored continuously and patched on a regular cycle. As demand grows, add CPU, memory and disk, or launch more servers.
What you get
Servers ready in minutes
Pick a ready Ubuntu, Debian, Arch Linux or Windows Server template and choose the CPU, memory and disk size. On Linux, your SSH key and first-boot script (cloud-init) are applied while the server comes up, so users and packages are already in place.
Servers that restart themselves after a hardware failure
When a physical host fails, the virtual servers on it restart automatically on other hosts. If a guest operating system hangs and stops responding, VM monitoring restarts it, so a server that freezes overnight doesn’t wait for someone to notice in the morning.
Fast, durable NVMe storage
Your databases and applications run on distributed, all-NVMe storage with no spinning-disk tier, so queries come back without waiting on the disk. Data is spread across several physical hosts, so losing one disk or one host never means losing data.
Resources that grow with you
Add CPU and memory from the portal, grow a disk or attach another one, with every request checked against your organisation’s quota. If hot add is turned on and the guest OS supports it, CPU and memory can be raised without powering the server down.
Share the load when traffic spikes
Put several web servers behind a load balancer ordered from the same portal and spread traffic across them. On campaign days or at month end, launching another server from a template and adding it to the pool takes minutes; delete it afterwards and the cost goes with it.
Design your own network
Isolated or internet-facing networks, DHCP, static routes, a DNS forwarder and an IPsec VPN to your office are set up from the portal in a few steps. Allocate and release your public IP addresses yourself, and expose only the service you mean to publish with a single rule.
A firewall that starts closed
Every organisation’s gateway firewall is handed over with a deny-all rule and logging on, so nothing gets in until you open it. UEFI Secure Boot can be enabled at boot time, so the server starts only with signed operating-system components.
Reach your server even when the network breaks
The browser console comes through the portal rather than the server’s own network, and every session opens with a one-time ticket. Even if a bad rule or IP setting cuts off SSH, you can sign in through the console and fix it yourself.
Native IPv6 and your own IP addresses
Nesnet networks run IPv4 and IPv6 side by side, and your IPv6 address comes from Nesnet’s own address block rather than a tunnel. When a tender or a customer specification asks for IPv6, there’s no extra project — request your own IPv6 prefix in the portal.
Keep critical servers on separate hardware
An anti-affinity rule places both nodes of a database or two web servers on different physical hosts, so a single hardware failure never takes out both. The opposite rule, keeping chatty servers on the same host, is available too.
Bring your own image and templates
Upload your own ISO or OVA to your organisation’s catalogue and build servers from it. Clone a server you’ve prepared or capture it as an organisation template, and power related servers on and off together as one application unit.
Licences that come with the server
Add Windows Server, SQL Server, Red Hat Enterprise Linux and SUSE licences to a server through rental licensing. Licences are rented monthly from Nesnet, and each one appears as its own line on the invoice.
Run the security appliances you already know
Virtual editions of widely used firewalls, routers, VPN servers and security-monitoring products are ready in the catalogue. Your network team keeps working in the interface it has used for years, and the appliance runs as a server on the same network as the rest of your fleet.
See who changed what
The portal’s audit trail shows who in your organisation changed which server or rule, and when. Give your team roles such as organisation administrator, catalogue author or console-only access, so opening a console for an intern doesn’t mean handing over the firewall rules.
A restore point before every update
Take a snapshot before a major update or configuration change, and roll back to that moment if things go wrong. A snapshot is not a backup; for lasting protection, the Backup product is added from the same portal.
Automate it
Anything you do in the portal can also be done through the platform API: creating servers, building networks, taking snapshots. Teams that script their test environments and delete them when they’re done get an identical configuration every single time.
GPUs for AI and graphics workloads
Add a GPU policy to the same kind of server to run AI inference, image processing or graphics workloads. Nesnet’s team builds a Windows Server template with the GPU driver already installed, so you skip the driver setup entirely.
Internet through two carriers, without interruption
Nesnet reaches the internet through two separate carriers, so traffic to and from your servers never depends on a single line. If one carrier’s line runs into trouble, traffic keeps flowing over the other.
Keep an eye on server health
Track CPU, memory, disk and network usage with the monitoring product and get a Slack or webhook alert when a threshold is crossed. Decide when to add capacity or resize based on data rather than guesswork, before a problem grows.
Take your servers with you
Package and download your servers in OVF/OVA, an open standard. If you ever move to another virtualisation environment, your server isn’t locked into a closed format, and the same package can be kept as an archive copy.
Why Nesnet
Your data stays in Türkiye
Your servers run in the Equinix Istanbul (Dudullu) and Türk Telekom Ankara data centres. You always know which country and which facility holds your data; it is never moved to a region abroad.
Nesnet’s own platform, local support
The same Nesnet team builds and runs the hardware, the network and the virtualisation layer; nothing is rented from another provider and resold. Your questions go to Nesnet’s support team, in Turkish or English.
Failure behaviour is measured, not assumed
Restart after a host failure and VM monitoring for hung guests are both switched on, and Nesnet verifies those settings against the live system. You know they’re active in production, not just on paper.
Changes ship step by step, with a way back
Upgrades and storage-layout changes are carried out one step at a time and verified after each. The tooling halts itself when it sees something unexpected, and the rollback path is written before the work starts.
Nesnet builds the templates
Linux and Windows Server templates are built, versioned and kept current by Nesnet’s team. Each Windows server gets a fresh SID at creation and an administrator password shown to you once, so two servers from the same template never collide in the domain.
e-Invoice and e-Archive
Invoices are issued in Türkiye as e-Invoice or e-Archive documents. Foreign payments drop off your finance team’s list.
Example scenarios
An online shop getting ready for sale week
Situation
An e-commerce company knows traffic will hit five times a normal day during its annual sale week. Last year the site ran on one server, slowed to a crawl in the first hour and left carts abandoned.
Outcome
Two days before the sale, three more web servers are launched from the same template and placed behind the load balancer. When the week ends the extra servers are deleted, and the invoice shows only the resources used that week.
A dev team that scripts its test environments
Situation
Before every release a development team needs a handful of test servers and waits two days each time. Because the servers are built by hand, no two come out the same.
Outcome
The team now creates servers through the API from the Ubuntu template and a first-boot script, and the wait drops from days to minutes. Every server boots identical, and when testing ends it is deleted and drops off the bill.
A company moving its out-of-warranty server room
Situation
A manufacturer’s accounting, ERP and file servers all go out of warranty in the same year. The replacement quote calls for a big upfront investment and a weekend of work.
Outcome
The servers are trialled on a copy first, then moved to Nesnet over a planned weekend; on Monday morning staff connect to the same applications. Hardware refreshes and warranty tracking disappear from the company’s budget lines.
Bringing your own firewall to the cloud
Situation
A company's network team has used the same firewall brand for years and has no wish to learn a new product for its cloud servers. Its rule sets and monitoring are built around that product.
Outcome
A virtual appliance of the same brand is launched from the catalogue and placed in front of the application servers. The team keeps working in the interface it knows, and the Nesnet gateway's default-deny rule stays in place as a second layer.
A company centralising its remote desktop
Situation
Staff across several branches use the same accounting application, installed separately on every PC. Licence tracking and updates are scattered.
Outcome
The application is installed once on a Windows server built from the RDS template, with per-user RDS licences rented from Nesnet. Staff connect over Remote Desktop, and updates happen on one server.
Who does what
| Nesnet | You |
|---|---|
| Runs the physical servers, disks and network links, and monitors temperature, power, fan and SSD sensors on every host. | Manage the operating system and applications inside the server, including their updates. |
| Builds and patches the virtualisation, storage, network and portal layers and keeps restart and balancing behaviour tuned. | Decide which ports to open and write your own firewall and NAT rules. |
| Builds and versions the Linux and Windows templates. | Keep the administrator password and SSH keys safe — the password is shown only once. |
| Operates internet connectivity, AS216146 routing and IPv6 announcements. | Assign portal roles to your team and review them regularly. |
| Hands over each gateway firewall with a default-deny policy. | Choose your backup plan: a snapshot is not a backup, and backup is a separate product. |
Who it is for
IT teams running their own applications
Teams that want full operating-system control for a web application, database, ERP or accounting software. Everything inside the server is yours; the hardware, storage and network underneath are Nesnet’s.
Teams that automate
Development teams that create and tear down test environments from scripts. Ready templates, a first-boot script and the platform API make the manual route unnecessary.
Network teams keeping the appliance they know
Network teams that want to run their familiar firewall brand in the cloud too. Virtual editions of widely used firewall and VPN appliances launch straight from the catalogue.
Companies working over remote desktop
Organisations whose branch staff reach the same application over remote desktop. The Remote Desktop Services (RDS) template and per-user licences come with the server.
Businesses with seasonal peaks
E-commerce, education and finance teams whose traffic or processing load multiplies on certain days of the year. Add capacity for the peak and hand it back afterwards, instead of buying hardware for your busiest day.
May not be the right fit
Teams that already ship containers and don't want to patch an operating system will find a virtual server is more work than they need. Nesly Serverless Platform or Managed Kubernetes is the better fit.
If the hardware must physically sit in your own building, this isn't the product. Nesnet can install the same platform in your data centre as a product-plus-consulting engagement.
If you plan to run dozens of servers with your own network design, identity provider and corporate circuit, look at Private Cloud rather than buying servers one at a time.
Under the hood
Your virtual server runs on a layered cloud platform that Nesnet built and operates itself. The cloud portal is where you create servers, build networks and open the console, and anything you can do in the portal you can also script against the API. The virtualisation layer balances load across physical hosts automatically, moves running servers away before maintenance and handles restarts after a failure. Distributed NVMe storage spreads data across several physical hosts, and in Nesnet’s own measurement write latency stayed below a millisecond. A software-defined network gives every organisation its own gateway and firewall. On top of that Nesnet adds its own work: it builds the templates, ships its own portal plug-ins, extends its own IPv6 network into customer networks and applies patches on a regular cycle.
Building blocks
- Cloud portal and API
- servers, networks and console on one screen; the same actions over the API
- Virtualisation layer
- restart after failure, fully automated load balancing, live migration
- Distributed NVMe storage
- data spread across several physical servers; one failure is not data loss
- Software-defined network
- a gateway, firewall, NAT, VPN and distributed routing for each organisation
- Load balancer
- ordered from the same portal, spreads traffic in front of your servers
How it compares
| Option | When it is the right choice |
|---|---|
| Your own hardware | If your load has been flat for years and you have a server room and people to look after it, owning the hardware can still make sense. The price is that spare machines and repairs are on you when something breaks. If the hardware has to stay in your building, Nesnet can install this same platform in your data centre. |
| Shared hosting | For a single website or a small shop managed through a control panel, with no need to touch the operating system, shared hosting is cheaper and simpler. Once you need your own software, your own services or Windows, you outgrow it quickly. |
| Global hyperscalers | If you need regions in several countries, hundreds of managed services or worldwide content delivery, a hyperscaler is the right call. If data residency in Türkiye, local e-invoicing and support in Turkish matter more, Nesnet is the better fit. |
| Nesnet Private Cloud | Once your server count grows, or you want your own network design, your own identity provider or your MPLS circuit brought into the cloud, move to Private Cloud. Your organisation gets a resource quota and you create as many servers inside it as you need. |
Technical specifications
| What you choose | vCPU, memory (GB RAM) and disk size, together with the template, in the portal |
|---|---|
| Resource model | Elastic resource pool; VM memory overhead is not charged to your quota |
| Storage | All-NVMe, distributed across several physical servers |
| Write latency | Below one millisecond (Nesnet measurement) |
| Disk options | Thin-provisioned, growable, extra, independent and shared disks |
| Clones and templates | Server clones, multi-server application units, your own template from a server |
| Linux templates | Ubuntu Server 22.04, 24.04, 26.04 · Debian 12 · Arch Linux |
|---|---|
| Windows templates | Windows Server 2016, 2022 (EN, TR, Datacenter), 2025 |
| Ready virtual appliances | Widely used firewall, router, VPN server and security-monitoring images |
| Your own images | ISO and OVA upload, organisation catalogue |
| Boot | UEFI Secure Boot, boot delay setting |
| Licensing | Windows Server, SQL Server, RHEL, SUSE — rental |
| First-boot configuration | cloud-init user-data and SSH key |
|---|---|
| Ready for SSH (Linux, measured) | About 2–3 minutes |
| Windows first boot | Fresh SID, password shown once |
| Console | HTML5 in the browser, one-time ticket |
| Console when the network breaks | Independent of the server's network |
| Inbound access | Public IP, DNAT for SSH 22 / RDP 3389 |
| Network services | Routed / isolated networks, DHCP, NAT, firewall, static routes, DNS forwarder, IPsec VPN |
| Gateway | Active-standby, distributed routing |
| Default firewall | Deny all, logging on |
| Autonomous system | AS216146, two separate upstream carriers |
| IPv6 | Native IPv6, dual-stack by default |
| Resizing | CPU and memory, quota-checked |
|---|---|
| Hot add | CPU and memory, when enabled and the guest OS supports it |
| Snapshot | Take, revert, delete (self-service) |
| Placement rules | Affinity / anti-affinity |
| Failure handling | Restart on another host after a failure, automatic balancing |
| Guest OS hangs | VM monitoring restarts it |
| Data protection | A single disk or host failure is not data loss |
| Audit trail | In the portal, per organisation |
| Portal roles | Organisation administrator, catalogue author, console-only access, Kubernetes cluster author |
| Password reset | A forgotten administrator password is reset from the portal |
| Portability | Download as OVF/OVA |
| Data centres | Equinix Istanbul (Dudullu), Türk Telekom Ankara |
If something breaks
A physical host fails
High availability detects the failure and restarts that host's virtual servers on other hosts in the cluster. To your server it looks like a reboot; its disk was already spread across other hosts, so the data is intact.
The operating system hangs
VM monitoring is on: if the guest stops responding, the VM is restarted. A Windows server that freezes at night doesn't wait for someone to notice and open the console.
A disk or a host is lost
The storage policy is built so that a single host or disk failure loses no data. While the failed part is replaced, your server keeps reading from the remaining copies.
A gateway node goes down
Your organisation's gateway runs active-standby; if the active node fails, the standby takes over with the same NAT and firewall rules.
You break your own network settings
The browser console doesn't depend on the server's network. If a wrong firewall rule or IP setting cuts off SSH, open the console and fix it.
A host needs maintenance
Maintenance is done one host at a time; virtual servers are moved live to another host before work on a host begins.
Frequently asked questions
Pricing and first steps
Pricing
Virtual servers are billed monthly for the resources you use. Line items are per vCPU, per GB of memory, per GB of NVMe storage and per public IP address. The price is set by the resources you use and the items you choose, and Nesnet’s team prepares your quote.
What the price depends on
- Number of vCPUs
- Memory (GB RAM)
- NVMe storage (GB)
- Number of public IP addresses
- Operating system and software licences: Windows Server per vCPU, SQL Server per vCPU, RDS per user
How to get started
- 1.Fill in the form or call 0800 606 1071.
- 2.Your needs are worked out in one call, with a sizing proposal and a price.
- 3.Your organisation is set up and you launch your first server from the portal.
Work out your server needs in one call
Tell Nesnet’s team how many servers, which operating systems and which licences you need; you’ll hear back within the same business day. Your server won’t depend on a single machine, and your data won’t leave Türkiye.