One node, or five hundred. Same firmware. Same file.
Start with a single box at the edge and grow to a multi-rack datacenter without re-architecting. The control plane lives in the nodes, so the cluster heals, rebalances, and fails over on its own.
From 1 to 500 nodes, no replatforming.
Add a node and run cluster add. Storage rebalances automatically, capacity and
compute grow together, and the price moves by exactly one node subscription.
- Grow without downtime. New nodes join the quorum and absorb load as they come online.
- Self-balancing storage. Data redistributes across the new fault domains on its own.
- Predictable cost. 500 nodes cost 500 node subscriptions — no core multiplier, no surprise at renewal.
When the edge drops, the datacenter takes over.
Edge sites stream data continuously to a central cluster via asynchronous replication. If an edge site fails, its workloads restart on the datacenter from the last replicated, signed restore point, with one-command failover and no data loss.
1. Stream
Edge sites replicate per-volume to the datacenter continuously and asynchronously.
2. Edge fails
A site goes dark. The datacenter detects the loss and confirms it through quorum.
3. Spin up
Workloads restart on the datacenter from the last signed restore point. One command, no data loss.
Bare metal to HA cluster in minutes.
Flash a node, run cluster create, add the next one whenever you need it.