AWS IoT Events Specialized2018年〜
A service that detects events from IoT sensor data and automatically executes actions
What It Does
AWS IoT Events is a service that monitors data from IoT devices and applications, and automatically executes actions when specific conditions or patterns are detected. You define state machines called detector models, and based on device state transitions, you can trigger actions such as sending SNS notifications, invoking Lambda functions, or sending messages to IoT Core.
Use Cases
Used for automatic alerts when factory equipment temperature exceeds a threshold, coordinated control of smart home devices based on state changes, anomaly detection and automatic shutdown of industrial equipment, and complex condition evaluation combining values from multiple sensors.
Everyday Analogy
Think of it like a factory monitoring operator. They constantly watch the instrument panel and automatically respond according to rules like 'sound the alarm if temperature exceeds 80 degrees' or 'close the valve if pressure drops.' They can also handle complex decisions that combine readings from multiple instruments.
What Is AWS IoT Events?
AWS IoT Events is a service that monitors telemetry data sent from IoT devices and detects and responds to events based on defined conditions. Beyond simple threshold monitoring, you can set up compound conditions combining data from multiple devices, as well as time-based conditions (such as when data hasn't arrived for a certain period). It automatically executes actions like notifications and device control in response to detected events, automating IoT system operations.
Detector Models
The core of IoT Events is the detector model. Defined as a state machine (state transition diagram), it lets you visually design each device state and the transition conditions between them. For example, you can define three states - 'Normal,' 'Warning,' and 'Critical' - and build a model where the state transitions to 'Warning' when temperature exceeds 70 degrees and to 'Critical' when it exceeds 80 degrees. You can configure actions for each state transition or while in a state, enabling flexible control based on conditions. For practical tips on using detector models, books (Amazon) are also a helpful resource.
Inputs and Actions
Data input to detector models can come via IoT Core's rule engine, direct submission to the IoT Events API, or integration from other AWS services. Available actions include sending notifications to SNS topics, invoking Lambda functions, sending messages to SQS queues, sending commands to devices through IoT Core, and writing data to DynamoDB. A single detector model can track multiple device instances individually, with independent state management for each device.
Things to Watch Out For
- Billed based on the number of message evaluations by detector models. Be mindful of costs when many devices send data at high frequency
- Complex state transition logic in detector models can be difficult to debug. Keep designs simple
- IoT Events specializes in event detection and immediate actions. For long-term data storage and trend analysis, use IoT Analytics alongside it
