Site·Edge·Federation·Disconnected
Kybernao runs on a node inside your boundary: a rack in a datacenter, a DIN rail in a substation, a pack on a patrol. Every node carries the full platform, so nothing depends on a link to work.
Choose by where the mission lives. The software, the controls and the evidence are identical.
For datacenters, plants and substations with stable power and space. One node can coordinate a fleet of others across a region.
For vehicles, vessels, forward teams and any contract where data is forbidden from leaving. The same platform, sized to carry.
Nodes share selected state through FedSpace, peer to peer. Kybernao Cloud can join as one more peer, hosted by us, and it is off by default. No node needs another node, or the cloud, to keep defending its own site.
You choose which data classes a node may share, and with which peers. The rest never leaves.
A node that loses its link queues updates and drains them in order when the link returns.
Each node keeps its own keys, ledger and policy. A compromised peer cannot command another node.
Disconnected operation is a normal mode, not a failure state. This is what a node does after the link drops.
The node notes the loss in its ledger and switches federation to queueing. Nothing else changes.
Sentinel was already running on the node, watching the routes Red has proven. There is no cloud to lose.
Guardian applies pre-approved containment inside the limits your operators set. No approval is waiting on a link.
Selected state and audit events forward to peers and your SIEM, signed and in sequence, with nothing lost.
Each node runs the complete platform. Pick the one that fits the place it will live.
NODE-S
TacticalRuns on battery or vehicle power for squads and forward teams.
NODE-R
RackFor datacenters and enterprise sites, and the coordinator for a fleet of nodes.
NODE-I
IndustrialPassive by default, for plants, substations and process environments.
NODE-X
ExpeditionaryFor austere sites where power is intermittent and support is days away.
Most teams start with one site, prove the loop on their own routes, then add nodes one environment at a time.
Rack it or rail it, connect identity and a span port, and start passive discovery.
Your mission engineer walks the first terrain graph and the routes Atlas resolved.
Red replays routes in the twin, Guardian containment is approved, and the first retests close findings.
New nodes join the fleet with the same policy, and federation is turned on only where you want it.
No. Every node runs the complete platform locally. A connection is only used for federation, SIEM forwarding and updates, and each of those is optional.
An optional hosted peer. A site can federate selected state to it the same way it would to another node, for fleet views across regions. It is off by default, and a node never depends on it to make a decision.
We provide signed update bundles that can be carried in on approved media. The node verifies the signature before it installs anything.
Talk to us about your environment. The NODE form factors are the supported path, and we can review other options with your mission engineer.
Most sites start with one. Larger plants add a NODE-I per process area, and a NODE-R can coordinate them all.
You can purge it on demand. The purge is recorded in the ledger, and the node's keys are revoked so it can no longer join the fleet.
We will size the hardware, map the network touch points and agree what, if anything, leaves the site.