Project Overview
A 24-hour AWS challenge to design and implement a production-grade, real-time IoT energy monitoring system. The project covers the full architecture from device data ingestion to time-series storage and analysis.
Architecture
- AWS IoT Core: Receives and authenticates MQTT messages from IoT energy sensors
- Amazon Kinesis Data Streams: Buffers and streams the incoming IoT data at scale
- AWS Lambda: Serverless functions process and transform each event in real time
- Amazon Timestream: Purpose-built time-series database stores processed energy metrics with efficient querying
- Terraform: All infrastructure provisioned as code for reproducibility and version control
Technical Decisions & Trade-offs
Timestream was chosen over DynamoDB or RDS for its native time-series optimizations (automatic data tiering, built-in interpolation). Kinesis was preferred over SQS for ordered, replayable streaming. Lambda kept costs low by removing the need for always-on compute.
Cost Optimization
The architecture was designed with cost in mind — serverless components scale to zero when idle, Timestream's magnetic store tier reduces long-term storage cost, and Kinesis shard allocation was right-sized for the expected throughput.
Outcome
A fully documented, IaC-driven real-time monitoring pipeline deployable in under 15 minutes, with sub-second latency from sensor event to stored time-series record.