VDS Hosting (KVM)

KVM (Kernel-based Virtual Machine) is full hardware virtualization. The VDS gets its own kernel, its own boot loader, its own virtual hardware. You can run Linux, Windows Server, FreeBSD or any custom ISO. LXC VPS is OS-level container virtualization sharing the host kernel: much lighter but Linux-only and no kernel-level changes.

Pick KVM VDS for Windows, kernel modules, custom kernels, or anything that needs hardware-level isolation; pick LXC VPS for cheaper Linux workloads where density and price-per-resource matter.

In our own DreamServer datacenter in Bucharest, Romania. EU jurisdiction, GDPR compliant. We are also a RIPE NCC LIR (AS57050) and operate the DSIX internet exchange, so the network path you get is the same one we run for our own production.

VDS instances are provisioned automatically within 10 minutes of payment. The customer panel is available immediately and gives you full self-service: power actions, backups, firewall rules, snapshots, an in-browser noVNC console (works for Windows, Linux, BSD, even before the network or SSH is up) and a one-click ISO reinstall.

Every KVM VDS uses an RBD volume in our Ceph distributed storage cluster with size=3. Each block of the virtual disk is split into placement groups, and CRUSH places three copies of every PG on separate OSDs across separate physical nodes. The cluster does not acknowledge a write until all three replicas are durable on disk, so you never lose data because of a single host or disk failure. With three replicas spread across three nodes, the cluster keeps serving reads and writes through the loss of two simultaneous nodes.

Yes. Because the VM disk lives in Ceph (not on a single host), we can live-migrate your VDS between hypervisors with zero downtime, no IP change and no reboot. This is how we patch hypervisors, replace failing servers and rebalance load while your Windows Server, Linux or BSD VDS stays online and serves traffic.

Live migration also means hardware maintenance is invisible to you: planned host evacuations, kernel updates and firmware upgrades happen during business hours with no service window required on your end.

The Ceph monitor quorum (Paxos) detects an OSD or full node failure within seconds and marks the affected OSDs out of the cluster map. CRUSH then rebalances the placement groups that lost a copy onto the remaining healthy OSDs, automatically rebuilding the missing replicas in the background. Your VDS continues to read and write the entire time, served by the surviving copies, and the rebuild rate is throttled so it does not impact production I/O.

For an extra off-cluster copy of your data, see Proxmox Backup Server. If you need single-tenant isolation on top of replicated storage, step up to a dedicated server.

Yes. KVM full virtualization supports any Windows Server edition (2019, 2022, 2025) with VirtIO drivers for performance. Bring your own license (BYOL) or arrange a SPLA license through us. The noVNC console works during installation just like a physical KVM-over-IP, including for the initial product key and partition setup.

Yes. The customer panel exposes a CD/DVD Disc Image File option where you can choose from our ISO library (AlmaLinux, CloudLinux, Debian, Proxmox VE, Rocky, Ubuntu Desktop and Server) or request a custom ISO upload.

Backups are an optional add-on, configured per VDS in the customer panel. Stored on a separate Proxmox Backup Server with client-side AES-256 encryption and content-defined deduplication. Standard retention is 7 daily, 4 weekly, 3 monthly. Pricing is per GB of backup space used.

Every VDS is covered by RTBH (Remotely Triggered Black Hole) filtering on our BGP uplinks, fed by sFlow telemetry on our edge routers. When a destination IP is flooded, the blackhole community is automatically announced to upstream transit providers so the attack is dropped at their network edge.

Effective against L3/L4 volumetric floods (UDP, SYN, amplification). Layer 7 attacks need application-level defenses inside your VDS.

Yes. Every VDS includes IPv6 in dual-stack with IPv4. Larger IPv6 allocations are available on request and we run our own IPv6 PI space as a RIPE NCC LIR.

Yes. CPU, RAM, SSD, IPv4, IPv6, backup space and max backups are all configurable from the customer panel and the changes take effect on next reboot. Disk can be expanded online; CPU and RAM hot-plug is supported on most distributions. Need a heavier setup? Move to a dedicated server; we handle the migration.

Power actions (start, reboot, stop, shut down), Backups + Backup Jobs + Schedules, Firewall + Firewall Options, Resource Graphs (CPU, Memory, Network, Disk I/O), one-click Reinstallation from the ISO library, Resource Notifications, Snapshots, Task History, VM Power Tasks and a full graphical noVNC Console. Read more about the managed administration bundle.

Out of the box: AlmaLinux 8.10 / 9.5 Minimal, CloudLinux 8.10 / 9.5, Debian 12.8 and 13.0, Proxmox VE 8.4.1 and 9.0.1, Rocky 8 and 9 Minimal, Ubuntu 20.04 / 22.04 / 24.04 in Desktop and Server variants. Anything else is a custom ISO upload, free of charge.

Default is 80 MB/s (640 Mbps). Steps up to 1 Gbps and uncapped (Unlimited) are available as a config option in the customer panel and reflected on the order form. Bandwidth is metered separately, with generous monthly quotas; uncapped traffic profiles are available on request.

Yes for both. The customer panel accepts an SSH public key at deploy and on every reinstall, and the ISO library supports custom uploads. The matching private key never leaves your machine. We never need or store your private key.

Yes, our NOC is on-call 24/7 for infrastructure incidents (datacenter, network, hypervisor, storage). Customer-application support during business hours is included; out-of-hours application-level support and escalation routing are part of the optional managed administration retainer.

Stripe (cards), bank transfer in EUR or RON, and on-account billing for established customers. Self-service ordering issues a pro-forma invoice; the final fiscal invoice is generated automatically on payment.

Yes. We invoice in EUR by default; RON invoices on request, converted at the BNR exchange rate of the invoice date. VAT is applied per Romanian / EU rules; reverse-charge VAT is available for valid VAT-registered EU customers outside Romania.

Yes. NDA is standard before any technical scoping. For workloads that touch personal data we sign a GDPR-aligned DPA. We are an EU-based company; data we process on your behalf stays in the EU unless you explicitly direct us otherwise.

99.9% monthly uptime for the VDS service, with proportional service credits for breach. Scheduled maintenance windows are announced via the customer panel and the status blog. Network-side redundancy is provided by multi-homed BGP transit on our LIR (AS57050) plus DSIX peering.

Yes. Common paths: P2V or V2V image import (qcow2, vmdk, vhdx), live rsync of a Linux file system to a fresh VDS, or a vzdump restore from another Proxmox cluster. The optional managed administration team handles the entire migration as a fixed-scope project, including a written runbook and acceptance test.

Yes. There is no minimum term. Cancellation takes effect at the end of the current billing cycle; we keep one final backup for 7 days after termination at no extra charge so you can pull data if needed. No exit fees, no clawbacks.

Yes. The VDS is yours; you get full root or Administrator inside the VM plus an in-browser noVNC console as an out-of-band fallback. We do not maintain a hidden management user inside the guest. Anything we run for managed customers is documented in your runbook in your repo.