Virtual Machines
Dedicated Linux VMs with fixed vCPU and memory, root access, attachable block disks, and predictable SKUs.
Pricing
Plans combine compute, memory, and additional disk space. VM compute SKUs are sized by vCPU and memory profile; additional disk space is priced by the gigabyte. Pick the compute tier first, then attach additional disk space to match your I/O requirements.
Compute
Loading VM SKUs…
Loading SKUs...
Full control,
no shared cores.
Teams choose VMs when they need a real machine boundary: kernel-level behaviour, legacy agents, licence-bound software, or lift-and-shift images that do not fit a lightweight container model.
Tower gives you dedicated vCPU and memory per instance, root access from first boot, attachable block disks, and the same predictable SKU model as our other compute lines—so capacity planning and procurement stay straightforward.
The trade-off is responsibility: guest OS patching, hardening, and application security remain yours. VMs also start slower and cost more per unit of elasticity than containers when you only need a stateless API.
Dedicated silicon,
replicated block disks.
Instances run on enterprise-grade AMD EPYC hardware with ECC memory. You are not sharing a kernel with strangers; noisy neighbours are kept out of your performance envelope by design.
Root and data disks live on our distributed block service—replicated, low-latency storage suited to databases and I/O-heavy workloads rather than a single point of failure on one machine.
Public and private connectivity options follow the same regional networking model as the rest of Tower, so VMs sit naturally beside databases, sites, and containerised apps in hybrid architectures.
What you get on Day One
Linux-first guests, SSH from boot, attachable disks, and integrated backup workflows—without running your own data centre.
Full root access
SSH in from first boot. Install packages, tune the kernel where your image allows, and bring supported custom images when your plan includes them. You own the guest; Tower owns the rack.
Flexible sizing
Choose vCPU and memory SKUs that match real workloads—from small dev boxes to larger production profiles—with clear limits so performance stays predictable.
Attachable block storage
Add data disks for databases, build caches, and log-heavy services. Disks persist across reboots and attach through the same Tower APIs you use elsewhere on the platform.
Automated backups
Snapshot and retention policies integrate with Tower backup services. Restore through the portal or API instead of manual disk copies.
Configuration & Platform Details
A concise reference for architecture reviews, procurement, and security questionnaires.
Platform specifications
Values reflect the Virtual Machines product line; exact limits may vary by region and account tier.
- Isolation model
- Dedicated virtual hardware per instance with fixed vCPU and memory allocations. No oversubscription of your purchased cores to other tenants.
- Compute platform
- AMD EPYC–based regions with ECC memory. Suitable for latency-sensitive and licence-bound workloads that need a full OS boundary.
- Persistent storage
- Replicated block volumes with online growth where the guest OS supports it. Designed for database and high-I/O use cases—not ephemeral scratch disks.
- Networking
- Regional public and private connectivity consistent with other Tower compute products. Interconnect options vary by plan and region—see pricing for bandwidth tiers.
- Operating systems
- Linux-first catalogue (common distributions and supported custom images). Windows Server is not offered as a standard managed image today—contact sales for bespoke requirements.
- Operational split
- Tower maintains the physical fleet, storage service, and platform APIs; customers maintain the guest OS, packages, and data inside the VM.
- Backups
- Integrated snapshot and retention policies with restore flows in the portal and API.