Skip to main content
Ballast

Broadcom's VMware licensing changes are pushing SMBs to look elsewhere

The SMB alternative to VMware and vCenter, built on Hyper-V.

Ballast is a control plane and distributed agent for Hyper-V hosts and failover clusters — a cohesive orchestration layer that does for Hyper-V what vCenter does for vSphere, minus the licensing model that just changed under you.

Community tier is free. No account required to read the docs.

A real capture from a running fabric, not a mockup.

Built to keep working when it's offline

The centre-dependency model works until the centre is the problem. Ballast's agent holds a durable, last-honoured copy of its own desired state and never invents intent — it just keeps enforcing what it was told, whether or not anything is watching.

Centralised model (vCenter)
Ballast
When the centre goes offline
Management is unavailable. Hosts keep running whatever they were last configured to run, but nothing new can be declared or corrected.
Each host's agent keeps enforcing its own last-honoured desired state — the same idempotent reconcile loop, just without a controller watching. The console shows it as Autonomous, not degraded.
Where desired state lives
Centrally, in the management plane. A host has no independent memory of its intended configuration.
Both centrally and cached on every host. Losing the centre never changes intent — only the ability to change it further.
Reconnecting after an outage
Manual reconciliation — comparing what happened while disconnected against what should be true.
Automatic. The agent reports its ObservedGeneration, drift is detected against the current desired state, and the centre resumes from there.
Settled Progressing Degraded AUTO Autonomous (centre offline) — deliberately not the "all is well" colour
A host settled at its current generation — reconcile steps, agent version and certificate status all reported per host, not inferred.

What it does

Build clusters from scratch

Declare hosts, storage and a witness, and Ballast coordinates with Windows Failover Clustering to bring the cluster up — no hand-run PowerShell sequence to keep in sync with reality.

Manage ROBO sites from one pane of glass

A remote-office host or a two-node cluster appears in the same fabric-wide host list as the main datacentre cluster — one table, not a separate console per site.

VM replication (Hyper-V Replica) to a standalone host or another cluster

A VM's replica shows up next to it, not in a separate tool — running/replicating state, target cluster and node, and which replicas have no live relationship and need attention.

DR runbooks

A runbook is the order your site comes back in: directory and DNS first, the databases that need them next, then the applications on top — one tier at a time, never forty VMs booting at once.

Templates

Capture a powered-off VM as a template — hardware profile, guest customisation and a generalised disk image — and deploy it repeatably anywhere the centre reaches.

VMware-to-Hyper-V migration

Register a VMware source, scan what it holds, and import a VM in a few clicks, with the console naming what needs attention — a raw device mapping, a blocked disk — rather than leaving you to work it out from a failed cmdlet.

Console access, from the browser

Every VM and every host has an interactive console reachable straight from the centre — no client to install, no separate jump box. A host's console is clearly marked as affecting every guest on it, not just one.

Pricing

Ballast is free to run on a single cluster, with no host limit and nothing held back. Additional clusters are licensed in packs that include vendor support. A deployment keeps the entitlement it was installed under for the life of the release line it was installed on — upgrading within a line never changes what you're entitled to.

Available now

Community

Free

  • One failover cluster of any size, plus unlimited standalone hosts
  • Full functionality, nothing held back
  • Active Directory sign-in for the console, with a local break-glass login
  • Desired-state reconciliation, drift detection and offline autonomy
  • VM replication (Hyper-V Replica) to a standalone host or another cluster
  • DR runbooks, VM templates, VMware migration
  • Fabric-wide view across every site you manage
  • Community support via GitHub issues and discussions
Get started on GitHub

In development

Commercial

Pricing TBC

  • Additional clusters, licensed in packs
  • SSO and SAML
  • Audit logging and compliance reporting
  • Vendor support with an SLA, included with every pack

What counts as a cluster?

Ballast Community includes one failover cluster (any size) plus unlimited standalone hosts. You only need a paid pack when you want to manage a second failover cluster or organisational features like SSO and audit logging.

Steady on its last-honoured state. Even offline.

Ballast is open-core, deliberately: the agent that runs on your hosts is open source and auditable, the centre is ours. Come look at the agent code, file an issue, or join the discussion.