L
18%
Licensing & Forkability
How cleanly the project remains free/open in practice: license clarity, anti-enclosure posture, and forkability if governance degrades.
Method
Final criteria and weights
Scores are assigned per criterion on a 0–100 scale and combined through a weighted composite. The model intentionally prioritizes licensing integrity, self-hosting, privacy posture, and composability, with lighter weighting on maturity because a forkable, self-hosted, high-integrity system can remain strategically valuable even when smaller.
S
18%
Self-Hosting & Offline Autonomy
How fully the stack can run on owned hardware without mandatory SaaS, with emphasis on offline behavior, ARM/SBC friendliness, and graceful degradation when disconnected.
P
14%
Privacy & Telemetry
Default data exhaust, telemetry, cloud assumptions, and how easily external calls can be removed or disabled.
G
12%
Governance & Capture Risk
Bus factor, maintainer concentration, foundation or corporate gravity, and long-run capture risk.
C
14%
Open Standards & Composability
Use of open standards and clean interfaces such as MQTT, LoRaWAN, OGC SensorThings, OGC services, REST, and other portable APIs.
X
10%
Operational Complexity & Attack Surface
Deployment weight, number of moving parts, difficulty of hardening, and total operational surface area.
D
9%
Domain Fitness (Agro / Env)
Direct fit for a sovereign agro / land / water / sensing stack rather than generic utility in the abstract.
M
5%
Maturity & Maintenance
Evidence of active releases, maintenance, documentation, and project viability.
Adversarial carry-through
Audit notes that changed or tightened the scoring
The updated ranking is not a copy of the first-pass ordering. These are the main corrections and hardening moves carried from the adversarial review into the final numbers.
Tania
Maintenance status upgraded after confirming ongoing activity and recent releases rather than assuming stagnation.
Plant-it
Reframed as an offline-first personal edge node rather than a server-centric application.
AgOpenGPS
Score held back by licensing ambiguity and version-specific legal clarity concerns despite very high field leverage.
SIP vs OpenSprinkler
SIP kept above OpenSprinkler on sovereignty because it avoids vendor-hardware anchoring and defaults to pure DIY control.
Node-RED
Privacy score deliberately penalized because telemetry and update checks exist and must be explicitly disabled in hardline local deployments.
OpenDroneMap / WebODM
Strong anti-capture licensing and unique aerial capability preserved, but complexity and compute weight were penalized more heavily.
GeoServer
Interop and geospatial leverage scored near the top, but attack surface and operational heft materially reduced the complexity score.
Matrix
Full scoring matrix
Composite score is the weighted sum of L, S, P, G, C, X, D, and M. Project names link to their detailed sections below.
| # | Project | L | S | P | G | C | X | D | M | Composite |
|---|---|---|---|---|---|---|---|---|---|---|
| 1 | Eclipse Mosquitto | 88 | 100 | 100 | 85 | 100 | 95 | 95 | 95 | 94.8 |
| 2 | FROST-Server | 92 | 95 | 98 | 85 | 100 | 80 | 95 | 90 | 92.6 |
| 3 | GeoServer | 95 | 93 | 95 | 90 | 100 | 70 | 95 | 95 | 92.2 |
| 4 | ChirpStack | 88 | 95 | 93 | 82 | 98 | 85 | 95 | 90 | 91.1 |
| 5 | OpenDroneMap / WebODM | 96 | 90 | 95 | 88 | 95 | 70 | 95 | 90 | 90.7 |
| 6 | SIP | 95 | 98 | 98 | 75 | 85 | 85 | 92 | 80 | 90.1 |
| 7 | farmOS + Field Kit | 95 | 90 | 90 | 88 | 85 | 75 | 98 | 90 | 89.2 |
| 8 | Node-RED | 88 | 95 | 80 | 80 | 100 | 80 | 90 | 95 | 88.6 |
| 9 | Tania | 88 | 93 | 93 | 80 | 88 | 85 | 92 | 80 | 88.3 |
| 10 | HortusFox | 88 | 93 | 90 | 78 | 80 | 85 | 80 | 85 | 85.7 |
| 11 | AgOpenGPS | 75 | 95 | 98 | 80 | 75 | 80 | 98 | 85 | 85.5 |
| 12 | OpenSprinkler | 88 | 90 | 82 | 75 | 80 | 85 | 90 | 90 | 84.8 |
| 13 | Plant-it | 95 | 95 | 98 | 78 | 40 | 90 | 70 | 85 | 82.4 |
LHow cleanly the project remains free/open in practice: license clarity, anti-enclosure posture, and forkability if governance degrades.
SHow fully the stack can run on owned hardware without mandatory SaaS, with emphasis on offline behavior, ARM/SBC friendliness, and graceful degradation when disconnected.
PDefault data exhaust, telemetry, cloud assumptions, and how easily external calls can be removed or disabled.
GBus factor, maintainer concentration, foundation or corporate gravity, and long-run capture risk.
CUse of open standards and clean interfaces such as MQTT, LoRaWAN, OGC SensorThings, OGC services, REST, and other portable APIs.
XDeployment weight, number of moving parts, difficulty of hardening, and total operational surface area.
DDirect fit for a sovereign agro / land / water / sensing stack rather than generic utility in the abstract.
MEvidence of active releases, maintenance, documentation, and project viability.
Ranking
Updated ranking by weighted composite
The highest-ranked entries are infrastructure-spine projects that combine low cloud dependence, portable standards, and strong local control. Domain-specific surfaces remain important, but they score lower when legal ambiguity, vendor anchoring, or narrower interoperability weaken the overall stack position.
Eclipse Mosquitto
Tiny, battle-tested MQTT broker with near-ideal self-hosting, privacy, composability, and low attack surface.
FROST-Server
Reference implementation of OGC SensorThings API and the cleanest structured sensor-data fabric in the set.
GeoServer
Open geospatial backbone with first-class OGC interoperability for maps, layers, features, and raster delivery.
ChirpStack
Private LoRaWAN network server with excellent MQTT and API integration for long-range field telemetry.
OpenDroneMap / WebODM
AGPL drone-mapping stack that converts imagery into orthomosaics, DEMs, point clouds, and 3D models offline.
SIP
DIY Raspberry Pi irrigation controller with strong sovereignty properties and no vendor hardware dependency.
farmOS + Field Kit
Most domain-specific farm record and planning platform in the set, with offline-capable field collection through Field Kit.
Node-RED
Best-in-class orchestration and glue layer, docked for telemetry and cloud-ecosystem gravity rather than technical weakness.
Tania
Lightweight open farm-management system with a cleaner operational profile than farmOS, but less ecosystem depth.
HortusFox
Self-hosted collaborative plant-management layer optimized for greenhouses, gardens, and smaller-scale plant environments.
AgOpenGPS
Open tractor-guidance and autosteer platform with huge real-world leverage, limited mainly by licensing ambiguity and weaker standards fit.
OpenSprinkler
Mature open irrigation controller ecosystem that remains practically strong but more vendor-anchored than SIP.
Plant-it
Offline-friendly GPL gardening companion with excellent personal privacy properties and minimal infrastructure relevance.
Project analyses
Detailed project-by-project analysis
Each section preserves the updated ranking logic while embedding official sites, documentation, repositories, and supporting project links directly alongside the analysis.
Rank #1
Eclipse Mosquitto 94.8
Tiny, battle-tested MQTT broker with near-ideal self-hosting, privacy, composability, and low attack surface.
L 88S 100P 100G 85C 100X 95D 95M 95
Why it scored here
Near-perfect on self-hosting, privacy, composability, and simplicity. The codebase is small, the deployment model is straightforward, and there is no structural cloud dependency. Governance under the Eclipse Foundation introduces only a modest capture concern because the broker is mature, widely adopted, and replaceable. In a sovereign agro stack, this is the default message bus and nervous system linking sensors, radio networks, controllers, and orchestration layers.
Stack role
Default message bus for ChirpStack uplinks, irrigation commands, automation flows, and sensor ingestion into higher layers.
Rank #2
FROST-Server 92.6
Reference implementation of OGC SensorThings API and the cleanest structured sensor-data fabric in the set.
L 92S 95P 98G 85C 100X 80D 95M 90
Why it scored here
Scores high because it formalizes sensor data around open OGC SensorThings semantics rather than ad hoc JSON or proprietary broker-side schemas. Self-hosting is clean, privacy posture is excellent, and portability across future stacks is unusually strong. Complexity is higher than the micro-infrastructure pieces because it is a Java plus database service, and governance inherits some institutional risk from Fraunhofer, but LGPL and reference-implementation status materially reduce enclosure risk.
Stack role
Canonical sensor ledger sitting above MQTT and LoRaWAN; stores structured observations, metadata, and tasking endpoints.
Rank #3
GeoServer 92.2
Open geospatial backbone with first-class OGC interoperability for maps, layers, features, and raster delivery.
L 95S 93P 95G 90C 100X 70D 95M 95
Why it scored here
Excels on licensing, interoperability, and domain leverage. It is the strongest geospatial publishing layer in the cohort and can serve drone products, land parcels, water infrastructure, environmental overlays, and field boundaries to desktop or web clients through mature OGC standards. The main penalty is operational complexity: the Java plus web stack is heavier than lightweight brokers or controllers and requires disciplined hardening, patching, and segmentation.
Stack role
Spatial backbone for orthomosaics, boundaries, hydrology, zoning, and sensor overlays.
Rank #4
ChirpStack 91.1
Private LoRaWAN network server with excellent MQTT and API integration for long-range field telemetry.
L 88S 95P 93G 82C 98X 85D 95M 90
Why it scored here
Almost mandatory wherever acreage, distance, and low-power sensing matter. It scores strongly on self-hosting, privacy, composability, and agricultural domain fit because it enables private LoRaWAN without forcing data through third-party network servers. The MIT license is clean but permissive, so proprietary wrapping remains possible. Governance also carries moderate capture risk because ecosystems around LoRaWAN often drift toward managed cloud control unless deliberately kept local.
Stack role
Long-range radio layer for remote nodes feeding local brokers and sensor ledgers.
Rank #5
OpenDroneMap / WebODM 90.7
AGPL drone-mapping stack that converts imagery into orthomosaics, DEMs, point clouds, and 3D models offline.
L 96S 90P 95G 88C 95X 70D 95M 90
Why it scored here
This remains strategically irreplaceable because nothing else in the list creates the same sovereign aerial vision layer. Licensing is exceptionally strong, outputs are standard geospatial artifacts, and local processing is fully viable. The main penalties are computational weight and operational complexity: image-processing pipelines are heavier, slower, and more infrastructure-intensive than message brokers, farm ledgers, or irrigation controllers.
Stack role
Aerial sensing and mapping engine feeding GeoServer and the broader spatial layer.
Rank #6
SIP 90.1
DIY Raspberry Pi irrigation controller with strong sovereignty properties and no vendor hardware dependency.
L 95S 98P 98G 75C 85X 85D 92M 80
Why it scored here
Scores unusually high on sovereignty because it combines GPL licensing, full local control, a Raspberry Pi deployment model, and no structural cloud requirement. It also avoids the hardware-vendor anchor present in commercial controller ecosystems. The ranking remains constrained by smaller project scale, lower bus factor, and a lighter ecosystem than the more mature infrastructure pieces.
Stack role
Local irrigation and actuation layer for valves, pumps, lights, and hydroponic components.
Rank #7
farmOS + Field Kit 89.2
Most domain-specific farm record and planning platform in the set, with offline-capable field collection through Field Kit.
L 95S 90P 90G 88C 85X 75D 98M 90
Why it scored here
Its main strength is domain seriousness: fields, crops, records, inputs, and planning are native concepts rather than generic entities. GPL licensing and self-hosting keep it aligned with sovereign deployment, and the Field Kit PWA preserves offline collection logic. The major drag is stack heaviness. Because it rides on a full Drupal-oriented web application layer, the operational burden and security surface are materially higher than in leaner Go or Python alternatives.
Stack role
Human-facing farm ledger for operations, records, and planning, linked to sensors and maps.
Rank #8
Node-RED 88.6
Best-in-class orchestration and glue layer, docked for telemetry and cloud-ecosystem gravity rather than technical weakness.
L 88S 95P 80G 80C 100X 80D 90M 95
Why it scored here
Nothing in the cohort matches its composability. It can bridge MQTT, HTTP, databases, device buses, dashboards, and external automation logic with minimal friction. The score drops because privacy is not perfect out of the box: telemetry and update checking exist and must be explicitly disabled in strict deployments. Governance and ecosystem culture also lean toward cloud and managed-service convenience, so sovereign installs should be local-only, tightly curated, and stripped of unnecessary internet-facing nodes.
Stack role
Automation and rule engine connecting data movement, actuation, dashboards, and payment logic.
Rank #9
Tania 88.3
Lightweight open farm-management system with a cleaner operational profile than farmOS, but less ecosystem depth.
L 88S 93P 93G 80C 88X 85D 92M 80
Why it scored here
The adversarial audit upgraded its maintenance status because the project remains active and viable rather than abandoned. Its design is leaner than farmOS, its self-hosting profile is strong, and privacy posture is straightforward. It does not match the standards depth of the infrastructure spine or the ecosystem maturity of farmOS, but it remains a credible management surface where simpler architecture matters more than platform breadth.
Stack role
Lean farm-management surface for areas, reservoirs, tasks, inventories, and crop cycles.
Rank #10
HortusFox 85.7
Self-hosted collaborative plant-management layer optimized for greenhouses, gardens, and smaller-scale plant environments.
L 88S 93P 90G 78C 80X 85D 80M 85
Why it scored here
HortusFox performs well as a personal or collaborative plant system, especially where journaling, treatments, plant inventories, and shared oversight matter. MIT licensing and straightforward self-hosting help. It ranks lower because standards integration is lighter, governance is maintainer-centered, and its domain fit is stronger at domestic or greenhouse scale than at broad field, machinery, or landscape scale.
Stack role
Collaborative plant care layer for greenhouses, gardens, and small distributed sites.
Rank #11
AgOpenGPS 85.5
Open tractor-guidance and autosteer platform with huge real-world leverage, limited mainly by licensing ambiguity and weaker standards fit.
L 75S 95P 98G 80C 75X 80D 98M 85
Why it scored here
If raw physical leverage were the only criterion, AgOpenGPS would rank much higher. It directly affects machine guidance, section control, and field execution, and it can run fully offline. The audit keeps it lower because license clarity remains weaker than the rest of the list and because integration patterns are more bespoke than the open-protocol spine built around MQTT, LoRaWAN, and OGC. The platform is strategically important but legally and architecturally less clean than the top tier.
Stack role
Machinery autonomy layer for tractor guidance, steering, and implement control.
Rank #12
OpenSprinkler 84.8
Mature open irrigation controller ecosystem that remains practically strong but more vendor-anchored than SIP.
L 88S 90P 82G 75C 80X 85D 90M 90
Why it scored here
OpenSprinkler stays respectable because the firmware is GPL, the hardware designs are open, and the community and release cadence remain healthy. It still ranks below SIP from a strict sovereignty perspective because the ecosystem is centered on a specific vendor and because external weather and cloud conveniences are more culturally baked into the product path. The right framing is not rejection but classification: strong and useful, yet more anchored to a commercial reference platform.
Stack role
Pragmatic irrigation controller for local scheduling, relay control, and field watering logic.
Rank #13
Plant-it 82.4
Offline-friendly GPL gardening companion with excellent personal privacy properties and minimal infrastructure relevance.
L 95S 95P 98G 78C 40X 90D 70M 85
Why it scored here
The audit upgraded Plant-it by recognizing its strong offline-first posture and local-device orientation. It remains last only because the scoring model privileges infrastructure, interoperability, and agro-stack composability. As a personal edge-node companion it is strong; as a central system in a sovereign farm architecture it is peripheral.
Stack role
Personal gardening and reminders app operating as an edge-node companion rather than a backbone service.
Architecture implication
Stack shape implied by the ranking
The resulting architecture naturally separates into a protocol and data spine, a spatial layer, a logic layer, and several domain-facing operational surfaces. The ranking does not erase the lower entries; it clarifies where each system belongs and what kind of sovereignty tradeoff accompanies it.
Signal spine
- Eclipse MosquittoDefault message bus
- ChirpStackPrivate LoRaWAN layer
- FROST-ServerStructured sensor ledger
Spatial layer
- OpenDroneMap / WebODMDrone processing and orthomosaic generation
- GeoServerOGC publishing and map serving
Logic layer
- Node-REDFlow engine, automation, and orchestration
Operational surfaces
- farmOS + Field KitFarm records, planning, and field data capture
- TaniaLean farm-management surface
- SIPDIY irrigation and actuation
- OpenSprinklerPragmatic open irrigation controller
- AgOpenGPSMachinery guidance and steering
- HortusFoxCollaborative greenhouse / garden management
- Plant-itPersonal edge-node plant companion