Core Concepts
Understanding how VeloxaHost models cloud infrastructure will help you make better architectural decisions from the start.
Instances
An instance is a single compute unit — either a system container (LXC) or a virtual machine (VM). Both types share the same API, console UI, and billing model.
| Property | System Container | Virtual Machine |
|---|---|---|
| Startup time | ~2 seconds | ~30–60 seconds |
| Kernel isolation | Shared host kernel | Own kernel |
| SSH access | ✅ | ✅ |
| Docker support | ✅ (nested) | ✅ |
| Snapshot speed | Instant (ZFS) | Seconds |
| Use case | Most workloads | Custom kernels, full isolation |
For most workloads, use system containers. They start faster, use less overhead, and behave identically to a full Linux server. Choose VMs when you need a custom kernel, GPU passthrough, or complete hardware isolation.
Projects
A project is your organizational namespace. All resources (instances, networks, storage, floating IPs) belong to a project. By default your account has one project. You can create multiple projects to separate environments (e.g., production, staging, dev) or teams.
- Resources in different projects are fully isolated from each other
- Billing rolls up across all projects to your account
- You can invite team members to specific projects with role-based access
Regions
A region is a physical data center location. All resources you create are bound to a region — an instance in vh-us-south-1 cannot directly communicate with an instance in vh-us-east-1 over private networking (only public IP). Choose your region closest to your end users for lowest latency. See Regions & Availability →
Networks
Every project gets a private network per region. Instances on the same private network can communicate over RFC-1918 addresses (10.x.x.x) without going through the public internet. See Networking guide →
Plans & Resources
Every instance is created from a plan that defines its CPU cores, memory, and root disk size. Plans are fixed — to resize, create a snapshot and launch a new instance from it at the desired plan size.
Start small — it's easy to scale up by snapshotting and re-launching. It's harder to scale storage down, so don't over-provision disk from the start.