Resilient mesh orchestration · Brisbane, Australia

The network will drop.
Nothing else should.

Altior Est keeps distributed fleets coordinating through dead zones, node failures, and live software upgrades — with zero confirmed work ever lost. It runs over the network you already own, and every claim it makes is enforced by an audited evidence ledger.

Evidence ledger — live fleet, 8 nodes, unedited

SOAK      7.0h  jobs=4188/4188 confirmed  elections=0  samples=85   → GREEN
KILL-SOAK 1.0h  jobs=599/599 confirmed   induced node kills=5       → GREEN
ROLLOUT   live upgrade, 8 nodes, mid-workload        jobs delta=+0GREEN
SOAK      early config error, honestly recorded                      → RED

The ledger can say no. It has refused runs of ours and kept the record — which is exactly why its GREEN entries mean something. Engineering teams are welcome to walk through the full ledger and methodology.

Two minutes, live, no cuts

Watch a fleet survive what site networks actually do

The coordinator's cable is pulled mid-operation. A node is killed and resurrected. A full software upgrade rolls through the live fleet mid-workload. Zero jobs lost.

4188 / 4188
jobs confirmed across a 7-hour overnight endurance run — zero elections, zero losses
142 / 0
mutants in the verification kill campaign / unexplained survivors
Δ +0
jobs lost during a live, zero-downtime software upgrade of all 8 nodes

What it does

Built for disconnected, degraded, intermittent, limited

Blast windows, dead zones behind benches, underground sections going dark, a nicked cable — the systems on top usually pay for it in dispatch gaps, telemetry holes, and manual resyncs. Altior Est makes them stop caring.

Coordination through dropouts

Work keeps being assigned, executed, and confirmed while parts of the network are dark. No confirmed work is ever lost — that is the contract, and the ledger enforces it.

Survives any node — including the leader

Kill any board mid-task and its work is re-placed automatically. Kill the coordinator and the fleet elects a new one and carries on. On camera, in the demo.

Dark sections resync themselves

A node that vanishes for a blast window returns, catches up, and rejoins — no manual resync, no operator intervention.

Zero-downtime upgrades

New software rolls through a live fleet node by node while work continues, gated by automated health checks that abort and roll back on any doubt.

Vendor-neutral by design

Runs over LTE, Wi-Fi, leaky feeder, Rajant kinetic mesh — whatever each site already owns. It's a software layer, not a rip-and-replace.

Your code stays yours

Workloads integrate as sealed programs behind a narrow interface. We never read your source; you never read ours. Security teams get the full architecture: per-device identity, signatures, encryption, forward secrecy, replay protection.

Why you can believe any of this

Verified like infrastructure, not like an app

Every release must pass the full gate before it can touch the fleet — and the results are recorded in an append-only evidence ledger your engineers can audit.

Model checking
The consensus core is checked exhaustively against a formal model — every reachable state, not a sample of them.
Conformance
The running implementation is tested against the verified model, so the code ships what the maths proved.
Mutation campaign
142 deliberate defects injected into the codebase; the test suite must catch them. Current score: zero unexplained survivors — every accepted mutant carries a written equivalence proof.
Evidence ledger
Soaks, kill-tests, and rollouts append GREEN or RED verdicts that cannot be quietly edited. RED entries stay on the record.

Engagement

A paid ten-week pilot, against your success criteria

No slideware phase, no open-ended trial. One site problem, written success criteria in the ledger's own metrics, ten weeks.

01

Scope

Your site's failure modes, your workloads, your network. Success criteria agreed in writing — measured by the ledger, not by opinion.

02

The gauntlet

Your topology mirrored; induced dropouts, node kills, coordinator kills, and a live upgrade run against your workloads. Weekly ledger report.

03

The verdict

Week ten: results against each criterion, in a report written for your management. Criteria green → an annual site licence conversation. Not green → you keep the evidence and we shake hands.

Talk to the founder

Altior Est is founder-led from Brisbane. Every enquiry is answered by the person who wrote the code.