AWS IoT Core ingestion pipeline, dual-path storage, and mobile-accessible dashboards for a multi-site grain and oilseed operation.
Stack
The Problem
A large grain and oilseed operation had IoT sensors logging data locally on farm equipment across multiple sites. Data was not backed up to the cloud, there was no cross-farm visibility, and equipment failures meant permanent data loss. The operations manager had no real-time view of what was happening across the portfolio. Alert conditions — grain temperature exceeding threshold, moisture levels outside range — were noticed only when someone was physically at the site or when a crop loss had already occurred.
The Result
Operations team has real-time cross-farm visibility from any device. Data loss from equipment failures eliminated. Client IT team operates and extends the infrastructure independently using the Terraform codebase and documentation we delivered.
What we built
Approach
All 40+ devices connect to AWS IoT Core using mutual TLS certificate authentication. Each device has a provisioned X.509 certificate and an IoT policy limiting it to publishing on its own device-scoped topic prefix. IoT Core's Rules Engine evaluates every incoming message and routes it to three simultaneous destinations. The hot path lands each reading in DynamoDB, keyed by device_id with timestamp as sort key. A TTL attribute prunes records older than 30 days automatically. The cold path routes to Kinesis Firehose, which batches and compresses readings into S3 as Parquet partitioned by site and date. Athena is configured with a Glue catalog table over the S3 prefix for ad-hoc historical queries. A Lambda function aggregates readings from the rules engine in one-minute windows and writes custom metrics to CloudWatch; those metrics power the dashboard and feed SNS alarm conditions for temperature and moisture thresholds. All infrastructure is defined in Terraform with remote state in S3 and DynamoDB table locking.
Full technical report
The full write-up includes architecture diagrams, technology selection rationale, implementation phases, and lessons learned. Available for qualified enquiries.