Tracking fleets in real time with AWS IoT Core and Amazon Location Service

This article presents a lightweight serverless architecture for real-time GPS fleet tracking using AWS IoT Core and Amazon Location Service, designed for small to medium-sized fleets as an alternative to enterprise solutions.

• GPS hardware connects to AWS IoT Core via MQTT with X.509 certificates; basic ingest topics route messages directly to IoT rules, bypassing the message broker
• Amazon Kinesis Data Streams buffers messages and enables batching before Lambda processes position updates to DynamoDB
• Amazon Location Service trackers evaluate positions against geofences and trigger EventBridge events when vehicles cross boundaries
• Two DynamoDB tables store current vehicle state and 24-hour position history; optional S3 archive handles longer-term retention
• API Gateway exposes REST APIs for queries and WebSocket APIs for real-time dashboard updates via DynamoDB streams
• Serverless approach suits GPS-focused tracking with basic telemetry for 20–200 vehicles at ~$0.85–$1.37 per vehicle monthly
• Apache Flink with Amazon MSK better serves large fleets requiring stateful processing, trip detection, and complex event correlation
• Position filtering and reporting intervals are the primary cost optimization levers; Amazon Location Service typically dominates bills without proximity-based filtering

The modular architecture allows scaling from serverless Kinesis+Lambda to managed Kafka+Flink while keeping device connectivity and MQTT topics unchanged.