System map

How the BioSwarm software is organized.

Twenty-six named software roles run the BioSwarm system — keeping the network coordinated, the readings trustworthy, the batteries fairly managed, and the records honest. Each role is a single, well-defined responsibility. Every signed record is publicly stamped with our root identity so partners can verify what we say without having to trust us.

Root identity40668c787c463ca5

Signature stampepoch26matrix2025

Design ratio1.618 (the golden ratio · same pattern leaves grow in)

Heartbeat1210 cycles per second

Target reliability99.78 %

How nodes agreemajority of nearby nodes confirm, with backup paths

Network shapeevery part can reach every other part, with backup routes

Self-correctingyes · industry-standard

Priority 1 · mission-critical Priority 2 · day-to-day backbone Priority 3 · monitoring + recovery Still to build · planned

Role 01 · priority 1 ready

AgriOrchestrator

also known as: Coordinator · coordination

The software that runs the recurring sync loop and keeps every part in step.

Role 02 · priority 1 ready

BatteryBridge

also known as: Compute bridge · compute link

The link between the live sensor stream and the heavier compute the demo runs on.

Role 03 · priority 2 ready

HiveRouter

also known as: Network router · message routing

The Bluetooth-mesh receiver that batches many sensors' readings into one signed upload.

Role 04 · priority 2 ready

SoilSignalCleaner

also known as: Reading cleaner · reading cleanup

The software that calibrates each sensor and flags readings that look wrong before they're stored.

Role 05 · priority 1 ready

LeafGuard

also known as: Hardware security · security

A chip-level signature on every reading so a tampered reading can't be slipped into the record.

Role 06 · priority 2 ready

BatteryPassportKeeper

also known as: Record keeper · record keeping

Per-node state — how the battery is doing, what the last reading was, when we last heard from it.

Role 07 · priority 2 ready

FloodConsensusDispatcher

also known as: Alert dispatcher · alerts

When several nearby sensors agree something is wrong, this is what sends the alert to the right place.

Role 08 · priority 3 ready

ZoneMemory

also known as: Fast memory · fast memory

A quick lookup so the dashboard loads instantly instead of recomputing everything from scratch.

Role 09 · priority 3 ready

EventDiary

also known as: History keeper · history keeping

An append-only diary of everything that's ever happened in a zone, so nothing is lost or rewritten.

Role 10 · priority 3 ready

HarvestMetrics

also known as: Measurements · measurements

Standard performance numbers — uptime, latency, throughput — published for any monitoring tool.

Role 11 · priority 2 ready

ColonyHealthMonitor

also known as: Health monitor · health monitoring

One overall "how is the swarm doing" score, built from five real signals (online, drift, …).

Role 12 · priority 2 ready

DutyCycleBalancer

also known as: Power budget · power budgeting

Asks each node to sample less often when its battery is low and more often when it has headroom.

Role 13 · priority 2 ready

WakeScheduler

also known as: Scheduler · scheduling

Spaces out radio wake-ups so a fleet of 200 nodes doesn't all transmit on the same millisecond.

Role 14 · priority 2 ready

BatteryBudgetAllocator

also known as: Resource allocator · resource allocation

Decides which sampling band each node belongs in based on its most recent battery reading.

Role 15 · priority 2 ready

TriageBotanist

also known as: Incident handler · incident handling

When something goes wrong, this ranks the issues and attaches a recommended next step for the operator.

Role 16 · priority 3 ready

ResprouterAgent

also known as: Retry handler · retries

The receiver's local queue — if the network drops, readings wait until it's safe to send again.

Role 17 · priority 2 ready

CircuitGuard

also known as: Trip protection · safe shutdown

If a firmware update starts failing on too many nodes, this pauses the rollout automatically.

Role 18 · priority 2 planned

BandwidthSteward

also known as: Traffic throttle · throttling

Per-customer traffic limits, so one site can't accidentally swamp the shared service.

Role 19 · priority 1 ready

ChipSigVerifier

also known as: Identity check · identity check

Verifies the chip-level signature on every reading before it's accepted into the record.

Role 20 · priority 1 ready

TenantWarden

also known as: Permissions · permissions

Keeps each customer's data and admin actions isolated; nothing crosses tenant lines.

Role 21 · priority 2 ready

ProofScribe

also known as: Audit-log keeper · audit logging

Produces the signed evidence bundles a third party (insurer, USDA, IBM) can verify on their own.

Role 22 · priority 2 ready

CalibrationSteward

also known as: Settings manager · settings

Stores per-node sensor calibration and the rules for what counts as an unusual reading.

Role 23 · priority 3 ready

PolicySwitchboard

also known as: Feature switches · feature toggles

Lets us turn behaviors on and off per zone or per node without reshipping software.

Role 24 · priority 2 ready

HiveSync

also known as: Sync handler · synchronization

Every 30 seconds, takes a signed snapshot of the zone's status and broadcasts it to listeners.

Role 25 · priority 3 ready

DailyBundler

also known as: Daily backup · backup

Once a day, ties together every signed record into a single tamper-evident summary you can archive.

Role 26 · priority 1 ready

ReplayHerbalist

also known as: Recovery handler · recovery

An offline auditor that can replay every signed record and verify the chain end-to-end without us.

25 of 26 roles ready · 1 still to build · Every signed record stamped under root identity 40668c787c463ca5