Tier 1 assumes continuous, automatic data upload from Pacific to the vendor cloud. Research indicates Cobalia currently syncs via Bluetooth only (Cobalia Sync mobile app). If there is no direct cloud upload path (SIM/WiFi from the Pacific unit itself), Tier 1 is not viable as described. This is the first question to clarify with OxyGuard.
Advantages
No edge hardware to buy, provision, or maintain
Fastest path to a working prototype
Lowest per-pond cost (sensors only)
Can demo with Parham's trout farm immediately
Limitations
Dependent on vendor's cloud uptime and API
No local intelligence beyond Pacific's built-in thresholds
Doesn't work if internet drops (data gap)
Locked to vendors that offer cloud APIs
Cannot combine multiple sensor brands locally
Latency: 1-5 minutes before data platform sees data
Tier 2 Details Edge Gateway
Advantages
Works on 2G (designed for poor connectivity)
Vendor-independent: any Modbus sensor works
Combines multiple sensor brands in one stream
Survives connectivity outages (data buffered)
Local alerting beyond Pacific's built-in thresholds
Remote management and SLA/service organisation needed
Additional point of failure (Pacific relay still works independently)
Tier 3 Details Edge AI Gateway
Advantages
Works on 2G (designed for poor connectivity)
Vendor-independent: any Modbus sensor works
Combines multiple sensor brands in one stream
Survives connectivity outages (data buffered)
Local alerting beyond Pacific's built-in thresholds
Real-time local intelligence (seconds, not minutes)
Predictive capability at the pond (prevent, not just react)
Competitive moat (nobody else does edge AI in aquaculture)
Limitations
Higher hardware cost per pond
Higher power draw (matters for off-grid farms)
Custom software development required
More complex provisioning (AI models on device)
Remote management and SLA/service organisation needed
Data Platform Architecture
Five layers. Every data point accessible via API. Extensible to any future data source (video, satellite imagery, manual input).
Integration Layer
Data-source-facing. Ingestion from edge gateways, vendor clouds, and external systems. Supports REST, MQTT, Flatfile, and Webhook. Edge gateways push compressed batches. Vendor clouds deliver via webhook or scheduled pull. Lab results and manual data via file upload.
Data Layer
All data resides here. Sensor readings, model results, business logic outcomes, manual input. Validated, transformed, and stored. Time-series database, streaming ingestion, blob storage, vector DB.
AI / Business Layer
Blue Aqua's proprietary aquaculture domain models. AI model deployment and management. Vector databases for knowledge retrieval. Business rules, alert logic, escalation workflows. This is where Blue Aqua's IP lives.
API Layer
Client-facing. Every data point and feature on the platform is accessible via REST APIs. Sensor readings, AI predictions, farm status, alerts, device health. AquaPilot and any future consumer (partner integrations, reporting tools) consumes data through this layer. No direct database access.
Compute Layer
Container orchestration. Auto-scaling. Manages deployments, load balancing, and infrastructure.
Why this matters: When Blue Aqua adds new data sources (underwater cameras, satellite imagery, weather APIs, feed tracking), they flow through the same pipeline. No new architecture needed.
Ingestion Protocols
Protocol
Source
Pattern
MQTT
Edge gateways (Tier 2 & 3)
Compressed batch every 5-15 min
HTTPS POST
Edge gateways (fallback), manual uploads
JSON payload, gzip compressed
Webhook
Vendor cloud APIs (Tier 1), external services
Event-driven, real-time
REST API pull
Vendor cloud APIs (Tier 1)
Scheduled polling every 1-5 min
Flat file
Manual data entry, CSV exports, lab results
Batch upload, validated on arrival
Edge Gateway
Every pond gets an edge device. It bridges sensors to the data platform and runs local intelligence.
Hardware & Software Requirements
Non-negotiable requirements for any edge device we deploy at pond level.
Category
Requirement
Deployment
Proprietary software pre-loaded. Standard provisioning process (warehouse flash).
Operation
Headless (no screen, no keyboard). IP65 enclosure for outdoor deployment.
Connectivity
Bluetooth and/or 2G to AquaPilot app. Compressed data sync to cloud.
Security
Full disk encryption. TLS for all communication. Per-device certificates. Remote wipe.
Updates
OTA software updates. Signed updates only. Auto-rollback on failure.
Management
Remotely accessible. Central fleet control. SSH via VPN only.
Questions for OxyGuard
Modbus TCP / Protocol
#
Question
Why we need it
1
Modbus TCP register map for Pacific Main Module
Required to read sensor values (DO, temp, salinity, pH, ORP) from our edge gateway. Without this we cannot integrate.
2
Register data types and scaling factors
Need to know integer vs float, scaling (e.g. value / 100 = actual reading), and byte order (big/little endian).
3
Polling frequency limits
How often can we poll registers without impacting Pacific performance? Target: every 10-30 seconds.
4
Master Box aggregation via Modbus
When using a Pacific Master Box (up to 20 probes), are all probe values accessible via the same Modbus TCP connection? Or does each probe require a separate connection?
5
Alarm/relay status registers
Can we read alarm thresholds and relay states via Modbus? Needed to sync safety rules between edge gateway and Pacific.
6
Simultaneous connections
Can our edge gateway and a local HMI/display both connect via Modbus TCP at the same time?
Cloud API (Cobalia)
#
Question
Why we need it
1
Cobalia availability and prerequisites
Is Cobalia included with Pacific hardware or a separate service agreement? Does OxyGuard need to provision cloud data collection per installation? What are the setup requirements?
2
Customer data consent
Does centralized data collection in Cobalia require explicit approval from end customers? Who owns the data? What are the data residency and privacy terms?
3
REST API availability
Does Cobalia expose a REST/HTTP API for programmatic data access? If yes, provide documentation. If not, is this on the roadmap?
4
Data export formats
Can we export data as JSON/CSV? Via API, webhook, or scheduled export?
5
Authentication method
API key, OAuth2, certificate-based? Per-device or per-account?
6
Data latency
How quickly after measurement does data appear in Cobalia? Currently sync is via Bluetooth (Cobalia Sync app). Is there a direct cloud upload path?
7
Historical data access
Can we query past readings via API? How far back? Rate limits?
8
Webhook / push notifications
Can Cobalia push data to our platform on new readings or alarm events?
9
API pricing
Is API access included with Cobalia subscription? Or separate licensing? Per-call or flat fee?
10
Multi-tenant / white-label
Can we operate Cobalia under Blue Aqua branding for our customers? Or is it OxyGuard-branded only?