Set up a new device to test V1 applications on V2 (Greengrass nucleus lite) - Amazon IoT Greengrass
Services or capabilities described in Amazon Web Services documentation might vary by Region. To see the differences applicable to the China Regions, see Getting Started with Amazon Web Services in China (PDF).

Set up a new device to test V1 applications on V2 (Greengrass nucleus lite)

Set up a new device with Greengrass nucleus lite to test generic components that you create to migrate your Amazon IoT Greengrass V1 applications to V2. Greengrass nucleus lite is a lightweight runtime optimized for resource-constrained devices. You can use this device to develop and test custom Greengrass components that run native processes. After you test your applications on a Greengrass nucleus lite device, you can deploy the V2 components to other devices running Greengrass nucleus lite or upgrade your existing V1 core devices to V2.

Step 1: Install Greengrass nucleus lite on a new device

Install the Greengrass nucleus lite on a new device. Follow the Greengrass nucleus lite installation guide to set up a device.

Note

Greengrass nucleus lite does not support local shadow service, client devices, or connectors now. Ensure your V1 applications don't rely on these features before proceeding with this guide.

Step 2: Create and deploy generic components to migrate Amazon IoT Greengrass V1 Lambda functions

To replicate the functionality of your Amazon IoT Greengrass V1 Lambda functions on Greengrass nucleus lite, you must convert them to generic components. This involves rewriting your Lambda function code to use the Amazon IoT Device SDK V2 or Amazon IoT Greengrass Component SDK instead of the Amazon IoT Greengrass Core SDK.

The following table lists the SDKs used in the V2 component examples in this guide:

Below examples show Lambda functions that use the Amazon IoT Greengrass V1 Core SDK and their equivalent generic components, with component code, recipes, and build instructions in multiple programming languages for two main scenarios:

  • Local communication - Components that communicate with other components on the same device using local pub/sub

  • Cloud communication - Components that communicate with Amazon IoT Core or other Amazon services

Scenario 1: Local Communication (publisher → processor → subscriber)

This scenario demonstrates converting a V1 Lambda function that uses local pub/sub communication to a V2 generic component.

Application architecture

This example involves three components:

  • Publisher Lambda publishes temperature data

  • Processor Lambda receives data and processes it

  • Processor Lambda publishes processed result to subscriber Lambda

The code example below focuses on the processor Lambda, which demonstrates both subscribing to and publishing messages for local communication.

V1 group subscriptions

In Amazon IoT Greengrass V1, the following group subscriptions enable communication between Lambda functions:

Subscription 1: publisher → processor

  • Source: Lambda (publisher)

  • Target: Lambda (processor)

  • Topic: sensors/temperature

Subscription 2: processor → subscriber

  • Source: Lambda (processor)

  • Target: Lambda (subscriber)

  • Topic: lambda/alerts

Processor Lambda function (V1)

Python
import greengrasssdk import json iot_client = greengrasssdk.client('iot-data') def lambda_handler(event, context): """ Receives temperature from publisher Lambda, processes it, and forwards to subscriber Lambda """ # Receive from publisher Lambda. sensor_id = event['sensor_id'] temperature = event['temperature'] print(f"Received from sensor {sensor_id}: {temperature}°F") # Process: Check if temperature is high. if temperature > 80: alert_data = { 'sensor_id': sensor_id, 'temperature': temperature, 'alert': 'HIGH_TEMPERATURE' } # Publish to another Lambda using greengrasssdk. iot_client.publish( topic='lambda/alerts', payload=json.dumps(alert_data) ) print(f"Alert sent to subscriber Lambda") return {'statusCode': 200}
Java
import com.amazonaws.greengrass.javasdk.IotDataClient; import com.amazonaws.greengrass.javasdk.model.PublishRequest; import com.amazonaws.services.lambda.runtime.Context; import com.google.gson.Gson; import java.nio.ByteBuffer; import java.nio.charset.StandardCharsets; import java.util.HashMap; import java.util.Map; public class TemperatureProcessorLambda { private static final Gson gson = new Gson(); private final IotDataClient iotDataClient; public TemperatureProcessorLambda() { this.iotDataClient = new IotDataClient(); } public String handleRequest(Map<String, Object> event, Context context) { /* * Receives temperature from publisher Lambda, * processes it, and forwards to subscriber Lambda */ // Receive from publisher Lambda. String sensorId = (String) event.get("sensor_id"); Number temp = (Number) event.get("temperature"); int temperature = temp.intValue(); System.out.println("Received from sensor " + sensorId + ": " + temperature + "°F"); // Process: Check if temperature is high. if (temperature > 80) { Map<String, Object> alertData = new HashMap<>(); alertData.put("sensor_id", sensorId); alertData.put("temperature", temperature); alertData.put("alert", "HIGH_TEMPERATURE"); // Publish to another Lambda using greengrasssdk. String payload = gson.toJson(alertData); PublishRequest publishRequest = new PublishRequest() .withTopic("lambda/alerts") .withPayload(ByteBuffer.wrap(payload.getBytes(StandardCharsets.UTF_8))); iotDataClient.publish(publishRequest); System.out.println("Alert sent to subscriber Lambda"); } return "Success"; } }
JavaScript
const greengrasssdk = require('aws-greengrass-core-sdk'); const iotClient = new greengrasssdk.IotData(); /** * Greengrass v1 Lambda function * Receives temperature from publisher Lambda, * processes it, and forwards to subscriber Lambda */ exports.handler = function(event, context) { // Receive from publisher Lambda. const sensorId = event.sensor_id; const temperature = event.temperature; console.log(`Received from sensor ${sensorId}: ${temperature}°F`); // Process: Check if temperature is high. if (temperature > 80) { const alertData = { sensor_id: sensorId, temperature: temperature, alert: 'HIGH_TEMPERATURE' }; // Publish to another Lambda using greengrasssdk. const params = { topic: 'lambda/alerts', payload: JSON.stringify(alertData) }; iotClient.publish(params, (err) => { if (err) { console.error('Error publishing alert:', err); context.fail(err); } else { console.log('Alert sent to subscriber Lambda'); context.succeed('Success'); } }); } else { context.succeed('Success'); } };
C
#include <aws/greengrass/greengrasssdk.h> #include <stdio.h> #include <string.h> #include <jansson.h> // For JSON parsing. static aws_greengrass_iot_data_client *iot_client = NULL; void on_message_received(const char *topic, const uint8_t *payload, size_t payload_len, void *user_data) { // Parse the incoming message. char *payload_str = strndup((char *)payload, payload_len); json_error_t error; json_t *event = json_loads(payload_str, 0, &error); free(payload_str); if (!event) { fprintf(stderr, "Error parsing JSON: %s\n", error.text); return; } // Receive from publisher Lambda. json_t *sensor_id_obj = json_object_get(event, "sensor_id"); json_t *temperature_obj = json_object_get(event, "temperature"); const char *sensor_id = json_string_value(sensor_id_obj); int temperature = json_integer_value(temperature_obj); printf("Received from sensor %s: %d°F\n", sensor_id, temperature); // Process: Check if temperature is high. if (temperature > 80) { // Create alert data. json_t *alert_data = json_object(); json_object_set_new(alert_data, "sensor_id", json_string(sensor_id)); json_object_set_new(alert_data, "temperature", json_integer(temperature)); json_object_set_new(alert_data, "alert", json_string("HIGH_TEMPERATURE")); // Convert to JSON string. char *alert_payload = json_dumps(alert_data, JSON_COMPACT); // Publish to another Lambda using greengrasssdk. aws_greengrass_publish_params params = { .topic = "lambda/alerts", .payload = (uint8_t *)alert_payload, .payload_len = strlen(alert_payload) }; aws_greengrass_iot_data_publish(iot_client, &params); printf("Alert sent to subscriber Lambda\n"); free(alert_payload); json_decref(alert_data); } json_decref(event); } int main(int argc, char *argv[]) { // Initialize Greengrass SDK. iot_client = aws_greengrass_iot_data_client_new(); // Subscribe to temperature sensor topic. aws_greengrass_subscribe_params subscribe_params = { .topic = "sensors/temperature", .callback = on_message_received, .user_data = NULL }; aws_greengrass_iot_data_subscribe(iot_client, &subscribe_params); printf("Temperature Processor Lambda started\n"); printf("Subscribed to sensors/temperature\n"); printf("Waiting for sensor data...\n"); // Keep the Lambda running. while (1) { sleep(1); } return 0; }
C++
#include <aws/greengrass/greengrasssdk.h> #include <iostream> #include <string> #include <memory> #include <jansson.h> // For JSON parsing. #include <unistd.h> class TemperatureProcessor { private: std::unique_ptr<aws_greengrass_iot_data_client, decltype(&aws_greengrass_iot_data_client_destroy)> iot_client; static void message_callback_wrapper(const char *topic, const uint8_t *payload, size_t payload_len, void *user_data) { auto* processor = static_cast<TemperatureProcessor*>(user_data); processor->on_message_received(topic, payload, payload_len); } public: TemperatureProcessor() : iot_client(aws_greengrass_iot_data_client_new(), aws_greengrass_iot_data_client_destroy) { if (!iot_client) { throw std::runtime_error("Failed to create Greengrass IoT client"); } } void on_message_received(const char *topic, const uint8_t *payload, size_t payload_len) { // Parse the incoming message. std::string payload_str(reinterpret_cast<const char*>(payload), payload_len); json_error_t error; json_t *event = json_loads(payload_str.c_str(), 0, &error); if (!event) { std::cerr << "Error parsing JSON: " << error.text << std::endl; return; } json_t *sensor_id_obj = json_object_get(event, "sensor_id"); json_t *temperature_obj = json_object_get(event, "temperature"); const char *sensor_id = json_string_value(sensor_id_obj); int temperature = json_integer_value(temperature_obj); std::cout << "Received from sensor " << sensor_id << ": " << temperature << "°F" << std::endl; if (temperature > 80) { send_alert(sensor_id, temperature); } json_decref(event); } void send_alert(const char *sensor_id, int temperature) { // Create alert data. json_t *alert_data = json_object(); json_object_set_new(alert_data, "sensor_id", json_string(sensor_id)); json_object_set_new(alert_data, "temperature", json_integer(temperature)); json_object_set_new(alert_data, "alert", json_string("HIGH_TEMPERATURE")); // Convert to JSON string. char *alert_payload = json_dumps(alert_data, JSON_COMPACT); // Publish to another Lambda using greengrasssdk. aws_greengrass_publish_params params = { .topic = "lambda/alerts", .payload = reinterpret_cast<uint8_t*>(alert_payload), .payload_len = strlen(alert_payload) }; aws_greengrass_iot_data_publish(iot_client.get(), &params); std::cout << "Alert sent to subscriber Lambda" << std::endl; free(alert_payload); json_decref(alert_data); } void subscribe_to_topic(const std::string& topic) { aws_greengrass_subscribe_params subscribe_params = { .topic = topic.c_str(), .callback = message_callback_wrapper, .user_data = this }; aws_greengrass_iot_data_subscribe(iot_client.get(), &subscribe_params); std::cout << "Temperature Processor Lambda started" << std::endl; std::cout << "Subscribed to " << topic << std::endl; std::cout << "Waiting for sensor data..." << std::endl; } void run() { // Keep the Lambda running. while (true) { sleep(1); } } }; int main(int argc, char *argv[]) { try { TemperatureProcessor processor; processor.subscribe_to_topic("sensors/temperature"); processor.run(); } catch (const std::exception& e) { std::cerr << "Error: " << e.what() << std::endl; return 1; } return 0; }

Generic component (V2)

To achieve the same functionality in Amazon IoT Greengrass V2, create a generic component with the following:

1. Component code
Python

Prerequisites: Before using this component code, install and verify the Amazon IoT Device SDK for Python on your Greengrass device:

# Install the SDK pip3 install awsiotsdk # Verify installation python3 -c "import awsiot.greengrasscoreipc.clientv2; print('SDK installed successfully')"

If you encounter dependency conflicts during installation, try installing a specific version of the Amazon IoT Device SDK.

If the verification command prints "SDK installed successfully", you're ready to use the component code:

from awsiot.greengrasscoreipc.clientv2 import GreengrassCoreIPCClientV2 from awsiot.greengrasscoreipc.model import ( PublishMessage, JsonMessage ) import time ipc_client = GreengrassCoreIPCClientV2() def on_sensor_data(event): """ Receives temperature from sensor publisher component, processes it, and forwards to alert component """ try: # Receive from publisher component. data = event.json_message.message sensor_id = data['sensor_id'] temperature = data['temperature'] print(f"Received from sensor {sensor_id}: {temperature}°F") # Process: Check if temperature is high. if temperature > 80: alert_data = { 'sensor_id': sensor_id, 'temperature': temperature, 'alert': 'HIGH_TEMPERATURE' } # Publish to another component (AlertHandler). ipc_client.publish_to_topic( topic='component/alerts', publish_message=PublishMessage( json_message=JsonMessage(message=alert_data) ) ) print(f"Alert sent to AlertHandler component") except Exception as e: print(f"Error processing sensor data: {e}") def main(): print("Temperature Processor component starting...") # Subscribe to sensor data from publisher component. ipc_client.subscribe_to_topic( topic='sensors/temperature', on_stream_event=on_sensor_data ) print("Subscribed to sensors/temperature") print("Waiting for sensor data...") # Keep running. while True: time.sleep(1) if __name__ == '__main__': main()
Java
import software.amazon.awssdk.aws.greengrass.GreengrassCoreIPCClientV2; import software.amazon.awssdk.aws.greengrass.model.PublishMessage; import software.amazon.awssdk.aws.greengrass.model.PublishToTopicRequest; import software.amazon.awssdk.aws.greengrass.model.JsonMessage; import software.amazon.awssdk.aws.greengrass.model.SubscribeToTopicRequest; import software.amazon.awssdk.aws.greengrass.model.SubscriptionResponseMessage; import java.util.HashMap; import java.util.Map; import java.util.Optional; public class TemperatureProcessor { private static GreengrassCoreIPCClientV2 ipcClient; public static void main(String[] args) { System.out.println("Temperature Processor component starting..."); try (GreengrassCoreIPCClientV2 client = GreengrassCoreIPCClientV2.builder().build()) { ipcClient = client; SubscribeToTopicRequest subscribeRequest = new SubscribeToTopicRequest() .withTopic("sensors/temperature"); ipcClient.subscribeToTopic( subscribeRequest, TemperatureProcessor::onSensorData, Optional.empty(), Optional.empty() ); System.out.println("Subscribed to sensors/temperature"); System.out.println("Waiting for sensor data..."); while (true) { Thread.sleep(1000); } } catch (Exception e) { System.err.println("Error: " + e.getMessage()); e.printStackTrace(); } } public static void onSensorData(SubscriptionResponseMessage message) { try { Map<String, Object> data = message.getJsonMessage().getMessage(); String sensorId = (String) data.get("sensor_id"); Number temp = (Number) data.get("temperature"); int temperature = temp.intValue(); System.out.println("Received from sensor " + sensorId + ": " + temperature + "F"); if (temperature > 80) { Map<String, Object> alertData = new HashMap<>(); alertData.put("sensor_id", sensorId); alertData.put("temperature", temperature); alertData.put("alert", "HIGH_TEMPERATURE"); JsonMessage jsonMessage = new JsonMessage().withMessage(alertData); PublishMessage publishMessage = new PublishMessage().withJsonMessage(jsonMessage); PublishToTopicRequest publishRequest = new PublishToTopicRequest() .withTopic("component/alerts") .withPublishMessage(publishMessage); ipcClient.publishToTopic(publishRequest); System.out.println("Alert sent to AlertHandler component"); } } catch (Exception e) { System.err.println("Error processing sensor data: " + e.getMessage()); } } }
JavaScript
const greengrasscoreipc = require('aws-iot-device-sdk-v2').greengrasscoreipc; class TemperatureProcessor { constructor() { this.ipcClient = null; } async start() { console.log('Temperature Processor component starting...'); try { this.ipcClient = greengrasscoreipc.createClient(); await this.ipcClient.connect(); const subscribeRequest = { topic: 'sensors/temperature' }; const streamingOperation = this.ipcClient.subscribeToTopic(subscribeRequest); streamingOperation.on('message', (message) => { this.onSensorData(message); }); streamingOperation.on('streamError', (error) => { console.error('Stream error:', error); }); streamingOperation.on('ended', () => { console.log('Subscription stream ended'); }); await streamingOperation.activate(); console.log('Subscribed to sensors/temperature'); console.log('Waiting for sensor data...'); } catch (error) { console.error('Error starting component:', error); process.exit(1); } } async onSensorData(message) { try { const data = message.jsonMessage.message; const sensorId = data.sensor_id; const temperature = data.temperature; console.log(`Received from sensor ${sensorId}: ${temperature}°F`); if (temperature > 80) { const alertData = { sensor_id: sensorId, temperature: temperature, alert: 'HIGH_TEMPERATURE' }; const publishRequest = { topic: 'component/alerts', publishMessage: { jsonMessage: { message: alertData } } }; await this.ipcClient.publishToTopic(publishRequest); console.log('Alert sent to AlertHandler component'); } } catch (error) { console.error('Error processing sensor data:', error); } } } // Start the component. const processor = new TemperatureProcessor(); processor.start();
C
#include <gg/buffer.h> #include <gg/error.h> #include <gg/ipc/client.h> #include <gg/map.h> #include <gg/object.h> #include <gg/sdk.h> #include <unistd.h> #include <stdbool.h> #include <stdio.h> #include <stdlib.h> #include <string.h> #include <pthread.h> #define SUBSCRIBE_TOPIC "sensors/temperature" #define PUBLISH_TOPIC "component/alerts" typedef struct { char sensor_id[64]; int64_t temperature; } AlertData; static pthread_mutex_t alert_mutex = PTHREAD_MUTEX_INITIALIZER; static pthread_cond_t alert_cond = PTHREAD_COND_INITIALIZER; static AlertData pending_alert; static bool has_pending_alert = false; static void *alert_publisher_thread(void *arg) { (void) arg; while (true) { pthread_mutex_lock(&alert_mutex); while (!has_pending_alert) { pthread_cond_wait(&alert_cond, &alert_mutex); } AlertData alert = pending_alert; has_pending_alert = false; pthread_mutex_unlock(&alert_mutex); GgBuffer sensor_id_buf = { .data = (uint8_t *)alert.sensor_id, .len = strlen(alert.sensor_id) }; GgMap payload = GG_MAP( gg_kv(GG_STR("sensor_id"), gg_obj_buf(sensor_id_buf)), gg_kv(GG_STR("temperature"), gg_obj_i64(alert.temperature)), gg_kv(GG_STR("alert"), gg_obj_buf(GG_STR("HIGH_TEMPERATURE"))) ); GgError ret = ggipc_publish_to_topic_json(GG_STR(PUBLISH_TOPIC), payload); if (ret != GG_ERR_OK) { fprintf(stderr, "Failed to publish alert\n"); } else { printf("Alert sent to AlertHandler component\n"); } } return NULL; } static void on_sensor_data( void *ctx, GgBuffer topic, GgObject payload, GgIpcSubscriptionHandle handle ) { (void) ctx; (void) topic; (void) handle; if (gg_obj_type(payload) != GG_TYPE_MAP) { fprintf(stderr, "Expected JSON message\n"); return; } GgMap map = gg_obj_into_map(payload); GgObject *sensor_id_obj; if (!gg_map_get(map, GG_STR("sensor_id"), &sensor_id_obj)) { fprintf(stderr, "Missing sensor_id field\n"); return; } GgBuffer sensor_id = gg_obj_into_buf(*sensor_id_obj); GgObject *temperature_obj; if (!gg_map_get(map, GG_STR("temperature"), &temperature_obj)) { fprintf(stderr, "Missing temperature field\n"); return; } int64_t temperature = gg_obj_into_i64(*temperature_obj); printf("Received from sensor %.*s: %lld°F\n", (int)sensor_id.len, sensor_id.data, (long long)temperature); if (temperature > 80) { pthread_mutex_lock(&alert_mutex); snprintf(pending_alert.sensor_id, sizeof(pending_alert.sensor_id), "%.*s", (int)sensor_id.len, sensor_id.data); pending_alert.temperature = temperature; has_pending_alert = true; pthread_cond_signal(&alert_cond); pthread_mutex_unlock(&alert_mutex); } } int main(void) { setvbuf(stdout, NULL, _IONBF, 0); printf("Temperature Processor component starting...\n"); gg_sdk_init(); GgError ret = ggipc_connect(); if (ret != GG_ERR_OK) { fprintf(stderr, "Failed to connect to Greengrass nucleus\n"); exit(1); } printf("Connected to Greengrass IPC\n"); // Start alert publisher thread. pthread_t alert_thread; if (pthread_create(&alert_thread, NULL, alert_publisher_thread, NULL) != 0) { fprintf(stderr, "Failed to create alert publisher thread\n"); exit(1); } GgIpcSubscriptionHandle handle; ret = ggipc_subscribe_to_topic( GG_STR(SUBSCRIBE_TOPIC), &on_sensor_data, NULL, &handle ); if (ret != GG_ERR_OK) { fprintf(stderr, "Failed to subscribe to topic\n"); exit(1); } printf("Subscribed to %s\n", SUBSCRIBE_TOPIC); printf("Waiting for sensor data...\n"); // Keep running. while (true) { sleep(1); } return 0; }
C++
#include <gg/ipc/client.hpp> #include <gg/buffer.hpp> #include <gg/object.hpp> #include <gg/types.hpp> #include <chrono> #include <condition_variable> #include <iostream> #include <mutex> #include <string> #include <string_view> #include <thread> struct AlertData { std::string sensor_id; int64_t temperature; }; static std::mutex alert_mutex; static std::condition_variable alert_cv; static AlertData pending_alert; static bool has_pending_alert = false; void alert_publisher_thread() { auto& client = gg::ipc::Client::get(); while (true) { std::unique_lock<std::mutex> lock(alert_mutex); alert_cv.wait(lock, [] { return has_pending_alert; }); AlertData alert = pending_alert; has_pending_alert = false; lock.unlock(); // Create alert payload as JSON string. std::string json_payload = "{\"sensor_id\":\"" + alert.sensor_id + "\",\"temperature\":" + std::to_string(alert.temperature) + ",\"alert\":\"HIGH_TEMPERATURE\"}"; // Convert to Buffer and publish. gg::Buffer buffer(json_payload); auto error = client.publish_to_topic("component/alerts", buffer); if (error) { std::cerr << "Failed to publish alert\n"; } else { std::cout << "Alert sent to AlertHandler component\n"; } } } class SensorCallback : public gg::ipc::LocalTopicCallback { void operator()( std::string_view topic, gg::Object payload, gg::ipc::Subscription& handle ) override { (void) topic; (void) handle; // Payload is a Buffer containing JSON string. if (payload.index() != GG_TYPE_BUF) { std::cerr << "Expected Buffer message\n"; return; } // Extract buffer using gg::get. auto buffer = gg::get<std::span<uint8_t>>(payload); std::string json_str(reinterpret_cast<const char*>(buffer.data()), buffer.size()); // Simple JSON parsing for demo. std::string sensor_id = "sensor1"; int64_t temperature = 0; // Extract temperature (simple string search). size_t temp_pos = json_str.find("\"temperature\":"); if (temp_pos != std::string::npos) { temp_pos += 14; // Skip "temperature". size_t end_pos = json_str.find_first_of(",}", temp_pos); if (end_pos != std::string::npos) { temperature = std::stoll(json_str.substr(temp_pos, end_pos - temp_pos)); } } std::cout << "Received from sensor " << sensor_id << ": " << temperature << "°F\n"; if (temperature > 80) { std::lock_guard<std::mutex> lock(alert_mutex); pending_alert = {sensor_id, temperature}; has_pending_alert = true; alert_cv.notify_one(); } } }; int main() { // Disable stdout buffering for real-time logging. std::cout.setf(std::ios::unitbuf); std::cout << "Temperature Processor component starting..." << std::endl; auto& client = gg::ipc::Client::get(); std::cout << "Got client instance" << std::endl; auto error = client.connect(); std::cout << "Connect returned, error code: " << error.value() << std::endl; if (error) { std::cerr << "Failed to connect to Greengrass nucleus: " << error.message() << std::endl; return 1; } std::cout << "Connected to Greengrass IPC" << std::endl; // Start alert publisher thread. std::thread alert_thread(alert_publisher_thread); alert_thread.detach(); // Handler must be static lifetime if subscription handle is not held. static SensorCallback handler; error = client.subscribe_to_topic("sensors/temperature", handler); if (error) { std::cerr << "Failed to subscribe to topic: " << error.message() << std::endl; return 1; } std::cout << "Subscribed to sensors/temperature" << std::endl; std::cout << "Waiting for sensor data..." << std::endl; // Keep running. while (true) { using namespace std::chrono_literals; std::this_thread::sleep_for(1s); } return 0; }
2. Build and package the component

Some languages require building or packaging before deployment.

Python

Python does not require compilation. The component can use the .py file directly.

Java

To build an executable JAR with all dependencies bundled:

  1. Create a pom.xml file in your project directory:

    <?xml version="1.0" encoding="UTF-8"?> <project xmlns="http://maven.apache.org/POM/4.0.0" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://maven.apache.org/POM/4.0.0 http://maven.apache.org/xsd/maven-4.0.0.xsd"> <modelVersion>4.0.0</modelVersion> <!-- Basic project information: organization, component name, and version --> <groupId>com.example</groupId> <artifactId>temperature-processor</artifactId> <version>1.0.0</version> <properties> <!-- Java 11 LTS (Long Term Support) is recommended for Greengrass v2 components --> <maven.compiler.source>11</maven.compiler.source> <maven.compiler.target>11</maven.compiler.target> </properties> <dependencies> <!-- AWS IoT Device SDK for Java - provides IPC client for Greengrass v2 local communication --> <dependency> <groupId>software.amazon.awssdk.iotdevicesdk</groupId> <artifactId>aws-iot-device-sdk</artifactId> <version>1.25.1</version> </dependency> </dependencies> <build> <plugins> <!-- Maven Shade Plugin - creates a standalone JAR with all dependencies included for Greengrass deployment --> <plugin> <groupId>org.apache.maven.plugins</groupId> <artifactId>maven-shade-plugin</artifactId> <version>3.2.4</version> <executions> <execution> <phase>package</phase> <goals> <goal>shade</goal> </goals> <configuration> <transformers> <!-- Set the main class for the executable JAR --> <transformer implementation="org.apache.maven.plugins.shade.resource.ManifestResourceTransformer"> <mainClass>TemperatureProcessor</mainClass> </transformer> </transformers> <filters> <!-- Exclude signature files to avoid security exceptions --> <filter> <artifact>*:*</artifact> <excludes> <exclude>META-INF/*.SF</exclude> <exclude>META-INF/*.DSA</exclude> <exclude>META-INF/*.RSA</exclude> </excludes> </filter> </filters> </configuration> </execution> </executions> </plugin> </plugins> </build> </project>
  2. Build the JAR:

    mvn clean package

    This creates target/temperature-processor-1.0.0.jar with all dependencies included.

  3. Upload the JAR to your S3 bucket for deployment.

JavaScript

To package your Node.js component with all dependencies:

  1. Create a package.json file:

    { "name": "temperature-processor", "version": "1.0.0", "description": "Temperature processor component for Greengrass v2", "main": "TemperatureProcessor.js", "dependencies": { "aws-iot-device-sdk-v2": "^1.21.0" }, "engines": { "node": ">=14.0.0" } }
  2. Install dependencies on your development machine:

    npm install

    This creates a node_modules folder containing the Amazon Amazon IoT Device SDK v2.

  3. Package for deployment:

    zip -r TemperatureProcessor.zip TemperatureProcessor.js node_modules/ package.json
  4. Upload the zip file to your S3 bucket for deployment.

Note

The Greengrass device must have Node.js runtime installed (version 14 or later). You don't need to run npm install on the Greengrass core device since the component artifact includes all dependencies in the bundled node_modules folder.

C

Prerequisites:

To build the SDK and the component, you will need the following build dependencies:

  • GCC or Clang

  • CMake (at least version 3.22)

  • Make or Ninja

Install build dependencies:

On Ubuntu/Debian:

sudo apt install build-essential cmake

On Amazon Linux:

sudo yum install gcc cmake make

Create a CMakeLists.txt file for your component:

cmake_minimum_required(VERSION 3.10) project(TemperatureProcessor C) set(CMAKE_C_STANDARD 11) # Add AWS Greengrass Component SDK add_subdirectory(aws-greengrass-component-sdk) # Build your component executable add_executable(temperature_processor temperature_processor.c) target_link_libraries(temperature_processor gg-sdk)

Build steps:

# Clone the AWS Greengrass Component SDK into your project git clone https://github.com/aws-greengrass/aws-greengrass-component-sdk.git # Build your component cmake -B build -D CMAKE_BUILD_TYPE=MinSizeRel make -C build -j$(nproc) # The binary 'temperature_processor' is in ./build/ # Upload this binary to S3 for deployment
C++

Prerequisites:

To build the SDK and the component, you will need the following build dependencies:

  • GCC or Clang with C++20 support

  • CMake (at least version 3.22)

  • Make or Ninja

Install build dependencies:

On Ubuntu/Debian:

sudo apt install build-essential cmake

On Amazon Linux:

sudo yum install gcc-c++ cmake make

Create a CMakeLists.txt file for your component:

cmake_minimum_required(VERSION 3.10) project(TemperatureProcessor CXX) set(CMAKE_CXX_STANDARD 20) # Add SDK as subdirectory add_subdirectory(aws-greengrass-component-sdk) # Add C++ SDK subdirectory add_subdirectory(aws-greengrass-component-sdk/cpp) add_executable(temperature_processor temperature_processor.cpp) target_link_libraries(temperature_processor gg-sdk++)

Build steps:

# Clone the AWS Greengrass Component SDK into your project git clone https://github.com/aws-greengrass/aws-greengrass-component-sdk.git # Build your component cmake -B build -D CMAKE_BUILD_TYPE=MinSizeRel make -C build -j$(nproc) # The binary 'temperature_processor' will be in ./build/ # Upload this binary to S3 for deployment
3. Component recipe

Update the "resources" array with the actual topics your component uses.

Python
{ "RecipeFormatVersion": "2020-01-25", "ComponentName": "com.example.TemperatureProcessor", "ComponentVersion": "1.0.0", "ComponentType": "aws.greengrass.generic", "ComponentDescription": "Receives sensor data and forwards alerts to AlertHandler", "ComponentPublisher": "[Your Company]", "ComponentConfiguration": { "DefaultConfiguration": { "accessControl": { "aws.greengrass.ipc.pubsub": { "com.example.TemperatureProcessor:pubsub:1": { "policyDescription": "Allows access to subscribe to sensor topics", "operations": [ "aws.greengrass#SubscribeToTopic" ], "resources": [ "sensors/temperature" ] }, "com.example.TemperatureProcessor:pubsub:2": { "policyDescription": "Allows access to publish to alert topics", "operations": [ "aws.greengrass#PublishToTopic" ], "resources": [ "component/alerts" ] } } } } }, "Manifests": [ { "Platform": { "os": "linux", "runtime": "*" }, "Lifecycle": { "run": "python3 -u {artifacts:path}/temperature_processor.py" }, "Artifacts": [ { "Uri": "s3://YOUR-BUCKET/artifacts/com.example.TemperatureProcessor/1.0.0/temperature_processor.py" } ] } ] }
Java
{ "RecipeFormatVersion": "2020-01-25", "ComponentName": "com.example.TemperatureProcessor", "ComponentVersion": "1.0.0", "ComponentType": "aws.greengrass.generic", "ComponentDescription": "Receives sensor data and forwards alerts to AlertHandler", "ComponentPublisher": "[Your Company]", "ComponentConfiguration": { "DefaultConfiguration": { "accessControl": { "aws.greengrass.ipc.pubsub": { "com.example.TemperatureProcessor:pubsub:1": { "policyDescription": "Allows access to subscribe to sensor topics", "operations": [ "aws.greengrass#SubscribeToTopic" ], "resources": [ "sensors/temperature" ] }, "com.example.TemperatureProcessor:pubsub:2": { "policyDescription": "Allows access to publish to alert topics", "operations": [ "aws.greengrass#PublishToTopic" ], "resources": [ "component/alerts" ] } } } } }, "Manifests": [ { "Platform": { "os": "linux", "runtime": "*" }, "Lifecycle": { "run": "java -jar {artifacts:path}/TemperatureProcessor.jar" }, "Artifacts": [ { "Uri": "s3://YOUR-BUCKET/artifacts/com.example.TemperatureProcessor/1.0.0/TemperatureProcessor.jar" } ] } ] }
JavaScript
{ "RecipeFormatVersion": "2020-01-25", "ComponentName": "com.example.TemperatureProcessor", "ComponentVersion": "1.0.0", "ComponentType": "aws.greengrass.generic", "ComponentDescription": "Receives sensor data and forwards alerts to AlertHandler", "ComponentPublisher": "[Your Company]", "ComponentConfiguration": { "DefaultConfiguration": { "accessControl": { "aws.greengrass.ipc.pubsub": { "com.example.TemperatureProcessor:pubsub:1": { "policyDescription": "Allows access to subscribe to sensor topics", "operations": [ "aws.greengrass#SubscribeToTopic" ], "resources": [ "sensors/temperature" ] }, "com.example.TemperatureProcessor:pubsub:2": { "policyDescription": "Allows access to publish to alert topics", "operations": [ "aws.greengrass#PublishToTopic" ], "resources": [ "component/alerts" ] } } } } }, "Manifests": [ { "Platform": { "os": "linux", "runtime": "*" }, "Lifecycle": { "run": "cd {artifacts:decompressedPath}/TemperatureProcessor && node TemperatureProcessor.js" }, "Artifacts": [ { "Uri": "s3://YOUR-BUCKET/artifacts/com.example.TemperatureProcessor/1.0.0/TemperatureProcessor.zip", "Unarchive": "ZIP" } ] } ] }
C/C++
{ "RecipeFormatVersion": "2020-01-25", "ComponentName": "com.example.TemperatureProcessor", "ComponentVersion": "1.0.0", "ComponentType": "aws.greengrass.generic", "ComponentDescription": "Receives sensor data and forwards alerts to AlertHandler", "ComponentPublisher": "[Your Company]", "ComponentConfiguration": { "DefaultConfiguration": { "accessControl": { "aws.greengrass.ipc.pubsub": { "com.example.TemperatureProcessor:pubsub:1": { "policyDescription": "Allows access to subscribe to sensor topics", "operations": [ "aws.greengrass#SubscribeToTopic" ], "resources": [ "sensors/temperature" ] }, "com.example.TemperatureProcessor:pubsub:2": { "policyDescription": "Allows access to publish to alert topics", "operations": [ "aws.greengrass#PublishToTopic" ], "resources": [ "component/alerts" ] } } } } }, "Manifests": [ { "Platform": { "os": "linux", "runtime": "*" }, "Lifecycle": { "run": "{artifacts:path}/temperature_processor" }, "Artifacts": [ { "Uri": "s3://YOUR-BUCKET/artifacts/com.example.TemperatureProcessor/1.0.0/temperature_processor" } ] } ] }

Scenario 2: Cloud Communication

This scenario demonstrates converting a V1 Lambda function that communicates with Amazon IoT Core to a V2 generic component.

Application architecture

This example uses a cloud-connected architecture:

  • Amazon IoT Core sends commands to the device

  • Controller Lambda receives commands and processes them

  • Controller Lambda sends telemetry data back to Amazon IoT Core

This example focuses on the controller Lambda, which demonstrates both receiving messages from and publishing messages to Amazon IoT Core.

V1 group subscriptions

In Amazon IoT Greengrass V1, the following group subscriptions enable communication between the Lambda function and Amazon IoT Core:

Subscription 1: IoT Cloud → Lambda

  • Source: IoT Cloud

  • Target: Lambda (DeviceController)

  • Topic: commands/device1

Subscription 2: Lambda → IoT Cloud

  • Source: Lambda (DeviceController)

  • Target: IoT Cloud

  • Topic: telemetry/device1

Controller Lambda function (V1)

Python
import greengrasssdk import json import time iot_client = greengrasssdk.client('iot-data') def lambda_handler(event, context): """ Receives commands from IoT Core, processes them, and sends telemetry back to cloud """ # Receive command from IoT Core. command = event.get('command') device_id = event.get('device_id', 'device1') print(f"Received command from cloud: {command}") # Process command. if command == 'get_status': status = get_device_status() # Send telemetry back to IoT Core. telemetry_data = { 'device_id': device_id, 'status': status, 'timestamp': time.time() } iot_client.publish( topic=f'telemetry/{device_id}', payload=json.dumps(telemetry_data) ) print(f"Telemetry sent to cloud: {telemetry_data}") return {'statusCode': 200} def get_device_status(): """Get current device status""" # Simulate getting device status. return 'online'
Java
import com.amazonaws.services.lambda.runtime.Context; import com.amazonaws.greengrass.javasdk.IotDataClient; import com.amazonaws.greengrass.javasdk.model.PublishRequest; import com.google.gson.Gson; import java.nio.ByteBuffer; import java.nio.charset.StandardCharsets; import java.util.HashMap; import java.util.Map; public class DeviceControllerLambda { private static final Gson gson = new Gson(); private final IotDataClient iotDataClient; public DeviceControllerLambda() { this.iotDataClient = new IotDataClient(); } public String handleRequest(Map<String, Object> event, Context context) { /* * Receives commands from IoT Core, * processes them, and sends telemetry back to cloud */ // Receive command from IoT Core. String command = (String) event.get("command"); String deviceId = event.containsKey("device_id") ? (String) event.get("device_id") : "device1"; System.out.println("Received command from cloud: " + command); // Process command. if ("get_status".equals(command)) { String status = getDeviceStatus(); // Send telemetry back to IoT Core. Map<String, Object> telemetryData = new HashMap<>(); telemetryData.put("device_id", deviceId); telemetryData.put("status", status); telemetryData.put("timestamp", System.currentTimeMillis() / 1000.0); String payload = gson.toJson(telemetryData); PublishRequest publishRequest = new PublishRequest() .withTopic("telemetry/" + deviceId) .withPayload(ByteBuffer.wrap(payload.getBytes(StandardCharsets.UTF_8))); iotDataClient.publish(publishRequest); System.out.println("Telemetry sent to cloud: " + telemetryData); } return "Success"; } private String getDeviceStatus() { // Simulate getting device status. return "online"; } }
JavaScript
const greengrasssdk = require('aws-greengrass-core-sdk'); const iotClient = new greengrasssdk.IotData(); /** * Receives commands from IoT Core and sends telemetry back. */ exports.handler = function(event, context) { console.log('Received command from IoT Core:', JSON.stringify(event)); const command = event.command; const deviceId = event.device_id || 'device1'; console.log(`Processing command: ${command}`); if (command === 'get_status') { const status = 'online'; const telemetryData = { device_id: deviceId, status: status, timestamp: Date.now() / 1000 }; // Publish telemetry to IoT Core using greengrasssdk. const params = { topic: `telemetry/${deviceId}`, payload: JSON.stringify(telemetryData) }; iotClient.publish(params, (err) => { if (err) { console.error('Error publishing telemetry:', err); context.fail(err); } else { console.log('Telemetry sent to IoT Core:', JSON.stringify(telemetryData)); context.succeed('Success'); } }); } else { context.succeed('Success'); } };
C
#include <aws/greengrass/greengrasssdk.h> #include <stdio.h> #include <string.h> #include <jansson.h> #include <time.h> static aws_greengrass_iot_data_client *iot_client = NULL; const char* get_device_status(void) { // Simulate getting device status. return "online"; } void on_cloud_command(const char *topic, const uint8_t *payload, size_t payload_len, void *user_data) { // Parse incoming command from IoT Core. char *payload_str = strndup((char *)payload, payload_len); json_error_t error; json_t *event = json_loads(payload_str, 0, &error); free(payload_str); if (!event) { fprintf(stderr, "Error parsing JSON: %s\n", error.text); return; } // Extract command and device_id. json_t *command_obj = json_object_get(event, "command"); json_t *device_id_obj = json_object_get(event, "device_id"); const char *command = json_string_value(command_obj); const char *device_id = device_id_obj ? json_string_value(device_id_obj) : "device1"; printf("Received command from cloud: %s\n", command); // Process command. if (command && strcmp(command, "get_status") == 0) { const char *status = get_device_status(); // Send telemetry back to IoT Core. json_t *telemetry_data = json_object(); json_object_set_new(telemetry_data, "device_id", json_string(device_id)); json_object_set_new(telemetry_data, "status", json_string(status)); json_object_set_new(telemetry_data, "timestamp", json_real(time(NULL))); char *telemetry_payload = json_dumps(telemetry_data, JSON_COMPACT); // Publish telemetry to IoT Core. char telemetry_topic[256]; snprintf(telemetry_topic, sizeof(telemetry_topic), "telemetry/%s", device_id); aws_greengrass_publish_params params = { .topic = telemetry_topic, .payload = (uint8_t *)telemetry_payload, .payload_len = strlen(telemetry_payload) }; aws_greengrass_iot_data_publish(iot_client, &params); printf("Telemetry sent to cloud: %s\n", telemetry_payload); free(telemetry_payload); json_decref(telemetry_data); } json_decref(event); } int main(int argc, char *argv[]) { // Initialize Greengrass SDK. iot_client = aws_greengrass_iot_data_client_new(); // Subscribe to commands from IoT Core. aws_greengrass_subscribe_params subscribe_params = { .topic = "commands/device1", .callback = on_cloud_command, .user_data = NULL }; aws_greengrass_iot_data_subscribe(iot_client, &subscribe_params); printf("Device Controller Lambda started\n"); printf("Subscribed to commands/device1\n"); printf("Waiting for commands from IoT Core...\n"); // Keep the Lambda running. while (1) { sleep(1); } return 0; }
C++
#include <aws/greengrass/greengrasssdk.h> #include <iostream> #include <string> #include <memory> #include <jansson.h> #include <ctime> #include <unistd.h> class DeviceController { private: std::unique_ptr<aws_greengrass_iot_data_client, decltype(&aws_greengrass_iot_data_client_destroy)> iot_client; static void message_callback_wrapper(const char *topic, const uint8_t *payload, size_t payload_len, void *user_data) { auto* controller = static_cast<DeviceController*>(user_data); controller->on_cloud_command(topic, payload, payload_len); } std::string get_device_status() { // Simulate getting device status. return "online"; } public: DeviceController() : iot_client(aws_greengrass_iot_data_client_new(), aws_greengrass_iot_data_client_destroy) { if (!iot_client) { throw std::runtime_error("Failed to create Greengrass IoT client"); } } void on_cloud_command(const char *topic, const uint8_t *payload, size_t payload_len) { // Parse incoming command from IoT Core. std::string payload_str(reinterpret_cast<const char*>(payload), payload_len); json_error_t error; json_t *event = json_loads(payload_str.c_str(), 0, &error); if (!event) { std::cerr << "Error parsing JSON: " << error.text << std::endl; return; } // Extract command and device_id. json_t *command_obj = json_object_get(event, "command"); json_t *device_id_obj = json_object_get(event, "device_id"); const char *command = json_string_value(command_obj); const char *device_id = device_id_obj ? json_string_value(device_id_obj) : "device1"; std::cout << "Received command from cloud: " << command << std::endl; // Process command. if (command && std::string(command) == "get_status") { std::string status = get_device_status(); // Send telemetry back to IoT Core. json_t *telemetry_data = json_object(); json_object_set_new(telemetry_data, "device_id", json_string(device_id)); json_object_set_new(telemetry_data, "status", json_string(status.c_str())); json_object_set_new(telemetry_data, "timestamp", json_real(std::time(nullptr))); char *telemetry_payload = json_dumps(telemetry_data, JSON_COMPACT); // Publish telemetry to IoT Core. std::string telemetry_topic = "telemetry/" + std::string(device_id); aws_greengrass_publish_params params = { .topic = telemetry_topic.c_str(), .payload = reinterpret_cast<uint8_t*>(telemetry_payload), .payload_len = strlen(telemetry_payload) }; aws_greengrass_iot_data_publish(iot_client.get(), &params); std::cout << "Telemetry sent to cloud: " << telemetry_payload << std::endl; free(telemetry_payload); json_decref(telemetry_data); } json_decref(event); } void subscribe_to_topic(const std::string& topic) { aws_greengrass_subscribe_params subscribe_params = { .topic = topic.c_str(), .callback = message_callback_wrapper, .user_data = this }; aws_greengrass_iot_data_subscribe(iot_client.get(), &subscribe_params); std::cout << "Device Controller Lambda started" << std::endl; std::cout << "Subscribed to " << topic << std::endl; std::cout << "Waiting for commands from IoT Core..." << std::endl; } void run() { // Keep the Lambda running. while (true) { sleep(1); } } }; int main(int argc, char *argv[]) { try { DeviceController controller; controller.subscribe_to_topic("commands/device1"); controller.run(); } catch (const std::exception& e) { std::cerr << "Error: " << e.what() << std::endl; return 1; } return 0; }

Generic component (V2)

To achieve the same functionality in Amazon IoT Greengrass V2, create a generic component with the following:

1. Component code
Python

Prerequisites: Before using this component code, install and verify the Amazon IoT Device SDK for Python on your Greengrass device:

# Install the SDK pip3 install awsiotsdk # Verify installation python3 -c "import awsiot.greengrasscoreipc.clientv2; print('SDK installed successfully')"

If you encounter dependency conflicts during installation, try installing a specific version of the Amazon IoT Device SDK.

If the verification command prints "SDK installed successfully", you're ready to use the component code:

from awsiot.greengrasscoreipc.clientv2 import GreengrassCoreIPCClientV2 from awsiot.greengrasscoreipc.model import QOS import json import time ipc_client = GreengrassCoreIPCClientV2() def on_command(event): """ Receives commands from IoT Core, processes them, and sends telemetry back to cloud """ try: # Receive command from IoT Core. data = json.loads(event.message.payload.decode('utf-8')) command = data.get('command') device_id = data.get('device_id', 'device1') print(f"Received command from cloud: {command}") # Process command. if command == 'get_status': status = get_device_status() # Send telemetry back to IoT Core. telemetry_data = { 'device_id': device_id, 'status': status, 'timestamp': time.time() } ipc_client.publish_to_iot_core( topic_name=f'telemetry/{device_id}', qos=QOS.AT_LEAST_ONCE, payload=json.dumps(telemetry_data).encode('utf-8') ) print(f"Telemetry sent to cloud: {telemetry_data}") except Exception as e: print(f"Error processing command: {e}") def get_device_status(): """Get current device status""" # Simulate getting device status. return 'online' def main(): print("Device Controller component starting...") # Subscribe to commands from IoT Core. ipc_client.subscribe_to_iot_core( topic_name='commands/device1', qos=QOS.AT_LEAST_ONCE, on_stream_event=on_command ) print("Subscribed to commands/device1 from IoT Core") print("Waiting for commands from cloud...") # Keep running. while True: time.sleep(1) if __name__ == '__main__': main()
Java
import software.amazon.awssdk.aws.greengrass.GreengrassCoreIPCClientV2; import software.amazon.awssdk.aws.greengrass.model.QOS; import software.amazon.awssdk.aws.greengrass.model.SubscribeToIoTCoreRequest; import software.amazon.awssdk.aws.greengrass.model.IoTCoreMessage; import software.amazon.awssdk.aws.greengrass.model.PublishToIoTCoreRequest; import java.nio.charset.StandardCharsets; import java.util.Optional; import java.util.regex.Pattern; import java.util.regex.Matcher; public class DeviceController { private static GreengrassCoreIPCClientV2 ipcClient; public static void main(String[] args) { System.out.println("Device Controller component starting..."); try (GreengrassCoreIPCClientV2 client = GreengrassCoreIPCClientV2.builder().build()) { ipcClient = client; SubscribeToIoTCoreRequest subscribeRequest = new SubscribeToIoTCoreRequest() .withTopicName("commands/device1") .withQos(QOS.AT_LEAST_ONCE); ipcClient.subscribeToIoTCore( subscribeRequest, DeviceController::onCommand, Optional.empty(), Optional.empty() ); System.out.println("Subscribed to commands/device1 from IoT Core"); System.out.println("Waiting for commands from cloud..."); while (true) { Thread.sleep(1000); } } catch (Exception e) { System.err.println("Error: " + e.getMessage()); e.printStackTrace(); } } public static void onCommand(IoTCoreMessage message) { try { String payload = new String(message.getMessage().getPayload(), StandardCharsets.UTF_8); // Simple JSON parsing. String command = extractJsonValue(payload, "command"); String deviceId = extractJsonValue(payload, "device_id"); if (deviceId == null || deviceId.isEmpty()) { deviceId = "device1"; } System.out.println("Received command from cloud: " + command); if ("get_status".equals(command)) { String status = getDeviceStatus(); // Build JSON manually. String telemetryJson = String.format( "{\"device_id\":\"%s\",\"status\":\"%s\",\"timestamp\":%.3f}", deviceId, status, System.currentTimeMillis() / 1000.0 ); byte[] telemetryBytes = telemetryJson.getBytes(StandardCharsets.UTF_8); PublishToIoTCoreRequest publishRequest = new PublishToIoTCoreRequest() .withTopicName("telemetry/" + deviceId) .withQos(QOS.AT_LEAST_ONCE) .withPayload(telemetryBytes); ipcClient.publishToIoTCore(publishRequest); System.out.println("Telemetry sent to cloud: " + telemetryJson); } } catch (Exception e) { System.err.println("Error processing command: " + e.getMessage()); } } private static String extractJsonValue(String json, String key) { Pattern pattern = Pattern.compile("\"" + Pattern.quote(key) + "\"\\s*:\\s*\"([^\"]+)\""); Matcher matcher = pattern.matcher(json); return matcher.find() ? matcher.group(1) : null; } private static String getDeviceStatus() { return "online"; } }
JavaScript
const greengrasscoreipc = require('aws-iot-device-sdk-v2').greengrasscoreipc; class DeviceController { constructor() { this.ipcClient = null; } async start() { console.log('Device Controller component starting...'); try { this.ipcClient = greengrasscoreipc.createClient(); await this.ipcClient.connect(); const subscribeRequest = { topicName: 'commands/device1', qos: 1 }; const streamingOperation = this.ipcClient.subscribeToIoTCore(subscribeRequest); streamingOperation.on('message', (message) => { this.onCommand(message); }); streamingOperation.on('streamError', (error) => { console.error('Stream error:', error); }); streamingOperation.on('ended', () => { console.log('Subscription stream ended'); }); await streamingOperation.activate(); console.log('Subscribed to commands/device1 from IoT Core'); console.log('Waiting for commands from cloud...'); } catch (error) { console.error('Error starting component:', error); process.exit(1); } } async onCommand(message) { try { const payload = message.message.payload.toString('utf-8'); const data = JSON.parse(payload); const command = data.command; const deviceId = data.device_id || 'device1'; console.log(`Received command from cloud: ${command}`); if (command === 'get_status') { const status = this.getDeviceStatus(); const telemetryData = { device_id: deviceId, status: status, timestamp: Date.now() / 1000 }; const telemetryJson = JSON.stringify(telemetryData); const publishRequest = { topicName: `telemetry/${deviceId}`, qos: 1, payload: Buffer.from(telemetryJson, 'utf-8') }; await this.ipcClient.publishToIoTCore(publishRequest); console.log(`Telemetry sent to cloud: ${telemetryJson}`); } } catch (error) { console.error('Error processing command:', error); } } getDeviceStatus() { return 'online'; } } // Start the component. const controller = new DeviceController(); controller.start();
C
#include <gg/buffer.h> #include <gg/error.h> #include <gg/ipc/client.h> #include <gg/map.h> #include <gg/object.h> #include <gg/sdk.h> #include <unistd.h> #include <stdbool.h> #include <stdio.h> #include <stdlib.h> #include <string.h> #include <time.h> #include <pthread.h> #define COMMAND_TOPIC "commands/device1" #define TELEMETRY_TOPIC "telemetry/device1" typedef struct { char device_id[64]; char command[64]; } CommandData; static pthread_mutex_t command_mutex = PTHREAD_MUTEX_INITIALIZER; static pthread_cond_t command_cond = PTHREAD_COND_INITIALIZER; static CommandData pending_command; static bool has_pending_command = false; const char* get_device_status(void) { // Simulate getting device status. return "online"; } static void *telemetry_publisher_thread(void *arg) { (void) arg; while (true) { pthread_mutex_lock(&command_mutex); while (!has_pending_command) { pthread_cond_wait(&command_cond, &command_mutex); } CommandData cmd = pending_command; has_pending_command = false; pthread_mutex_unlock(&command_mutex); // Process command. if (strcmp(cmd.command, "get_status") == 0) { const char *status = get_device_status(); // Create telemetry JSON string. char telemetry_json[512]; snprintf(telemetry_json, sizeof(telemetry_json), "{\"device_id\":\"%s\",\"status\":\"%s\",\"timestamp\":%ld}", cmd.device_id, status, time(NULL)); GgBuffer telemetry_buf = { .data = (uint8_t *)telemetry_json, .len = strlen(telemetry_json) }; // Publish telemetry to IoT Core. GgError ret = ggipc_publish_to_iot_core(GG_STR(TELEMETRY_TOPIC), telemetry_buf, 0); if (ret != GG_ERR_OK) { fprintf(stderr, "Failed to publish telemetry to IoT Core\n"); } else { printf("Telemetry sent to cloud: device_id=%s, status=%s\n", cmd.device_id, status); } } } return NULL; } static void on_cloud_command( void *ctx, GgBuffer topic, GgBuffer payload, GgIpcSubscriptionHandle handle ) { (void) ctx; (void) topic; (void) handle; printf("Received command from IoT Core: %.*s\n", (int)payload.len, payload.data); // Parse JSON payload (comes as raw buffer from IoT Core). // For simplicity, we'll do basic string parsing. // Extract command and device_id from JSON string. char payload_str[512]; snprintf(payload_str, sizeof(payload_str), "%.*s", (int)payload.len, payload.data); // Simple JSON parsing (looking for "command":"get_status"). char *command_start = strstr(payload_str, "\"command\""); char *device_id_start = strstr(payload_str, "\"device_id\""); if (command_start) { pthread_mutex_lock(&command_mutex); char *cmd_value = strstr(command_start, ":"); if (cmd_value) { cmd_value = strchr(cmd_value, '"'); if (cmd_value) { cmd_value++; char *cmd_end = strchr(cmd_value, '"'); if (cmd_end) { size_t cmd_len = cmd_end - cmd_value; if (cmd_len < sizeof(pending_command.command)) { strncpy(pending_command.command, cmd_value, cmd_len); pending_command.command[cmd_len] = '\0'; } } } } // Extract device_id or use default. if (device_id_start) { char *dev_value = strstr(device_id_start, ":"); if (dev_value) { dev_value = strchr(dev_value, '"'); if (dev_value) { dev_value++; char *dev_end = strchr(dev_value, '"'); if (dev_end) { size_t dev_len = dev_end - dev_value; if (dev_len < sizeof(pending_command.device_id)) { strncpy(pending_command.device_id, dev_value, dev_len); pending_command.device_id[dev_len] = '\0'; } } } } } else { strcpy(pending_command.device_id, "device1"); } printf("Received command from cloud: %s\n", pending_command.command); has_pending_command = true; pthread_cond_signal(&command_cond); pthread_mutex_unlock(&command_mutex); } } int main(void) { setvbuf(stdout, NULL, _IONBF, 0); printf("Device Controller component starting...\n"); gg_sdk_init(); GgError ret = ggipc_connect(); if (ret != GG_ERR_OK) { fprintf(stderr, "Failed to connect to Greengrass nucleus\n"); exit(1); } printf("Connected to Greengrass IPC\n"); // Start telemetry publisher thread. pthread_t telemetry_thread; if (pthread_create(&telemetry_thread, NULL, telemetry_publisher_thread, NULL) != 0) { fprintf(stderr, "Failed to create telemetry publisher thread\n"); exit(1); } // Subscribe to commands from IoT Core. GgIpcSubscriptionHandle handle; ret = ggipc_subscribe_to_iot_core( GG_STR(COMMAND_TOPIC), 0, &on_cloud_command, NULL, &handle ); if (ret != GG_ERR_OK) { fprintf(stderr, "Failed to subscribe to IoT Core topic\n"); exit(1); } printf("Subscribed to %s\n", COMMAND_TOPIC); printf("Waiting for commands from IoT Core...\n"); while (true) { sleep(1); } return 0; }
C++
#include <gg/ipc/client.hpp> #include <gg/buffer.hpp> #include <gg/object.hpp> #include <gg/types.hpp> #include <chrono> #include <condition_variable> #include <ctime> #include <iostream> #include <mutex> #include <string> #include <string_view> #include <thread> constexpr std::string_view COMMAND_TOPIC = "commands/device1"; constexpr std::string_view TELEMETRY_TOPIC = "telemetry/device1"; struct CommandData { std::string device_id; std::string command; }; static std::mutex command_mutex; static std::condition_variable command_cv; static CommandData pending_command; static bool has_pending_command = false; std::string get_device_status() { // Simulate getting device status. return "online"; } void telemetry_publisher_thread() { auto& client = gg::ipc::Client::get(); while (true) { std::unique_lock<std::mutex> lock(command_mutex); command_cv.wait(lock, [] { return has_pending_command; }); CommandData cmd = pending_command; has_pending_command = false; lock.unlock(); // Process command. if (cmd.command == "get_status") { std::string status = get_device_status(); // Get current timestamp. auto now = std::time(nullptr); // Create telemetry JSON payload. std::string telemetry_payload = "{\"device_id\":\"" + cmd.device_id + "\",\"status\":\"" + status + "\",\"timestamp\":" + std::to_string(now) + "}"; // Publish telemetry to IoT Core. gg::Buffer telemetry_buffer(telemetry_payload); auto error = client.publish_to_iot_core(TELEMETRY_TOPIC, telemetry_buffer); if (error) { std::cerr << "Failed to publish telemetry to IoT Core: " << error.message() << std::endl; } else { std::cout << "Telemetry sent to cloud: device_id=" << cmd.device_id << ", status=" << status << std::endl; } } } } class CloudCommandCallback : public gg::ipc::IoTCoreTopicCallback { void operator()( std::string_view topic, gg::Object payload, gg::ipc::Subscription& handle ) override { (void) topic; (void) handle; // Payload is a Buffer containing JSON string from IoT Core. if (payload.index() != GG_TYPE_BUF) { std::cerr << "Expected Buffer message\n"; return; } // Extract buffer. auto buffer = gg::get<std::span<uint8_t>>(payload); std::string json_str(reinterpret_cast<const char*>(buffer.data()), buffer.size()); std::cout << "Received command from IoT Core: " << json_str << std::endl; // Simple JSON parsing for demo. std::string command; std::string device_id = "device1"; // Default // Extract command. size_t cmd_pos = json_str.find("\"command\":"); if (cmd_pos != std::string::npos) { size_t start = json_str.find("\"", cmd_pos + 10) + 1; size_t end = json_str.find("\"", start); if (end != std::string::npos) { command = json_str.substr(start, end - start); } } // Extract device_id if present. size_t dev_pos = json_str.find("\"device_id\":"); if (dev_pos != std::string::npos) { size_t start = json_str.find("\"", dev_pos + 12) + 1; size_t end = json_str.find("\"", start); if (end != std::string::npos) { device_id = json_str.substr(start, end - start); } } if (!command.empty()) { std::lock_guard<std::mutex> lock(command_mutex); pending_command = {device_id, command}; has_pending_command = true; command_cv.notify_one(); std::cout << "Received command from cloud: " << command << std::endl; } } }; int main() { // Disable stdout buffering for real-time logging in systemd/Greengrass. std::cout.setf(std::ios::unitbuf); std::cout << "Device Controller component starting..." << std::endl; auto& client = gg::ipc::Client::get(); auto error = client.connect(); if (error) { std::cerr << "Failed to connect to Greengrass nucleus: " << error.message() << std::endl; return 1; } std::cout << "Connected to Greengrass IPC" << std::endl; // Start telemetry publisher thread. std::thread telemetry_thread(telemetry_publisher_thread); telemetry_thread.detach(); // Subscribe to commands from IoT Core. static CloudCommandCallback handler; error = client.subscribe_to_iot_core(COMMAND_TOPIC, handler); if (error) { std::cerr << "Failed to subscribe to IoT Core topic: " << error.message() << std::endl; return 1; } std::cout << "Subscribed to " << COMMAND_TOPIC << std::endl; std::cout << "Waiting for commands from IoT Core..." << std::endl; // Keep running. while (true) { using namespace std::chrono_literals; std::this_thread::sleep_for(1s); } return 0; }
2. Build and package the component

Some languages require building or packaging before deployment.

Python

Python does not require compilation. The component can use the .py file directly.

Java

To build an executable JAR with all dependencies bundled:

  1. Create a pom.xml file in your project directory:

    <?xml version="1.0" encoding="UTF-8"?> <project xmlns="http://maven.apache.org/POM/4.0.0" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://maven.apache.org/POM/4.0.0 http://maven.apache.org/xsd/maven-4.0.0.xsd"> <modelVersion>4.0.0</modelVersion> <!-- Basic project information: organization, component name, and version --> <groupId>com.example</groupId> <artifactId>device-controller</artifactId> <version>1.0.0</version> <properties> <!-- Java 11 LTS (Long Term Support) is recommended for Greengrass v2 components --> <maven.compiler.source>11</maven.compiler.source> <maven.compiler.target>11</maven.compiler.target> </properties> <dependencies> <!-- AWS IoT Device SDK for Java - provides IPC client for Greengrass v2 cloud communication --> <dependency> <groupId>software.amazon.awssdk.iotdevicesdk</groupId> <artifactId>aws-iot-device-sdk</artifactId> <version>1.25.1</version> </dependency> </dependencies> <build> <plugins> <!-- Maven Shade Plugin - creates a standalone JAR with all dependencies included for Greengrass deployment --> <plugin> <groupId>org.apache.maven.plugins</groupId> <artifactId>maven-shade-plugin</artifactId> <version>3.2.4</version> <executions> <execution> <phase>package</phase> <goals> <goal>shade</goal> </goals> <configuration> <transformers> <!-- Set the main class for the executable JAR --> <transformer implementation="org.apache.maven.plugins.shade.resource.ManifestResourceTransformer"> <mainClass>DeviceController</mainClass> </transformer> </transformers> <filters> <!-- Exclude signature files to avoid security exceptions --> <filter> <artifact>*:*</artifact> <excludes> <exclude>META-INF/*.SF</exclude> <exclude>META-INF/*.DSA</exclude> <exclude>META-INF/*.RSA</exclude> </excludes> </filter> </filters> </configuration> </execution> </executions> </plugin> </plugins> </build> </project>
  2. Build the JAR:

    mvn clean package

    This creates target/device-controller-1.0.0.jar with all dependencies included.

  3. Upload the JAR to your S3 bucket for deployment.

JavaScript

To package your Node.js component with all dependencies:

  1. Create a package.json file:

    { "name": "device-controller", "version": "1.0.0", "description": "Device controller component for Greengrass v2", "main": "DeviceController.js", "dependencies": { "aws-iot-device-sdk-v2": "^1.21.0" }, "engines": { "node": ">=14.0.0" } }
  2. Install dependencies on your development machine:

    npm install

    This creates a node_modules folder containing the Amazon Amazon IoT Device SDK v2.

  3. Package for deployment:

    zip -r DeviceController.zip DeviceController.js node_modules/ package.json
  4. Upload the zip file to your S3 bucket for deployment.

Note

The Greengrass device must have Node.js runtime installed (version 14 or later). You don't need to run npm install on the Greengrass core device since the component artifact includes all dependencies in the bundled node_modules folder.

C

Prerequisites:

To build the SDK and the component, you will need the following build dependencies:

  • GCC or Clang

  • CMake (at least version 3.22)

  • Make or Ninja

Install build dependencies:

On Ubuntu/Debian:

sudo apt install build-essential cmake

On Amazon Linux:

sudo yum install gcc cmake make

Create a CMakeLists.txt file for your component:

cmake_minimum_required(VERSION 3.10) project(DeviceController C) set(CMAKE_C_STANDARD 11) # Add AWS Greengrass Component SDK add_subdirectory(aws-greengrass-component-sdk) # Build your component executable add_executable(device_controller device_controller.c) target_link_libraries(device_controller gg-sdk)

Build steps:

# Clone the AWS Greengrass Component SDK into your project git clone https://github.com/aws-greengrass/aws-greengrass-component-sdk.git # Build your component cmake -B build -D CMAKE_BUILD_TYPE=MinSizeRel make -C build -j$(nproc) # The binary 'device_controller' is in ./build/ # Upload this binary to S3 for deployment
C++

Prerequisites:

To build the SDK and the component, you will need the following build dependencies:

  • GCC or Clang with C++20 support

  • CMake (at least version 3.22)

  • Make or Ninja

Install build dependencies:

On Ubuntu/Debian:

sudo apt install build-essential cmake

On Amazon Linux:

sudo yum install gcc-c++ cmake make

Create a CMakeLists.txt file for your component:

cmake_minimum_required(VERSION 3.10) project(DeviceController CXX) set(CMAKE_CXX_STANDARD 20) # Add SDK as subdirectory add_subdirectory(aws-greengrass-component-sdk) # Add C++ SDK subdirectory add_subdirectory(aws-greengrass-component-sdk/cpp) add_executable(device_controller device_controller.cpp) target_link_libraries(device_controller gg-sdk++)

Build steps:

# Clone the AWS Greengrass Component SDK into your project git clone https://github.com/aws-greengrass/aws-greengrass-component-sdk.git # Build your component cmake -B build -D CMAKE_BUILD_TYPE=MinSizeRel make -C build -j$(nproc) # The binary 'device_controller' will be in ./build/ # Upload this binary to S3 for deployment
3. Component recipe

Update the "resources" array with the actual topics your component uses.

Python
{ "RecipeFormatVersion": "2020-01-25", "ComponentName": "com.example.DeviceController", "ComponentVersion": "1.0.0", "ComponentType": "aws.greengrass.generic", "ComponentDescription": "Receives commands from IoT Core and sends telemetry back to cloud", "ComponentPublisher": "[Your Company]", "ComponentConfiguration": { "DefaultConfiguration": { "accessControl": { "aws.greengrass.ipc.mqttproxy": { "com.example.DeviceController:mqttproxy:1": { "policyDescription": "Allows access to subscribe to IoT Core topics", "operations": [ "aws.greengrass#SubscribeToIoTCore" ], "resources": [ "commands/device1" ] }, "com.example.DeviceController:mqttproxy:2": { "policyDescription": "Allows access to publish to IoT Core topics", "operations": [ "aws.greengrass#PublishToIoTCore" ], "resources": [ "telemetry/device1" ] } } } } }, "ComponentDependencies": { "aws.greengrass.TokenExchangeService": { "VersionRequirement": ">=2.0.0", "DependencyType": "HARD" } }, "Manifests": [ { "Platform": { "os": "linux", "runtime": "*" }, "Lifecycle": { "run": "python3 -u {artifacts:path}/device_controller.py" }, "Artifacts": [ { "Uri": "s3://YOUR-BUCKET/artifacts/com.example.DeviceController/1.0.0/device_controller.py" } ] } ] }
Java
{ "RecipeFormatVersion": "2020-01-25", "ComponentName": "com.example.DeviceController", "ComponentVersion": "1.0.0", "ComponentType": "aws.greengrass.generic", "ComponentDescription": "Receives commands from IoT Core and sends telemetry back to cloud", "ComponentPublisher": "[Your Company]", "ComponentConfiguration": { "DefaultConfiguration": { "accessControl": { "aws.greengrass.ipc.mqttproxy": { "com.example.DeviceController:mqttproxy:1": { "policyDescription": "Allows access to subscribe to IoT Core topics", "operations": [ "aws.greengrass#SubscribeToIoTCore" ], "resources": [ "commands/device1" ] }, "com.example.DeviceController:mqttproxy:2": { "policyDescription": "Allows access to publish to IoT Core topics", "operations": [ "aws.greengrass#PublishToIoTCore" ], "resources": [ "telemetry/device1" ] } } } } }, "ComponentDependencies": { "aws.greengrass.TokenExchangeService": { "VersionRequirement": ">=2.0.0", "DependencyType": "HARD" } }, "Manifests": [ { "Platform": { "os": "linux", "runtime": "*" }, "Lifecycle": { "run": "java -jar {artifacts:path}/DeviceController.jar" }, "Artifacts": [ { "Uri": "s3://YOUR-BUCKET/artifacts/com.example.DeviceController/1.0.0/DeviceController.jar" } ] } ] }
JavaScript
{ "RecipeFormatVersion": "2020-01-25", "ComponentName": "com.example.DeviceController", "ComponentVersion": "1.0.0", "ComponentType": "aws.greengrass.generic", "ComponentDescription": "Receives commands from IoT Core and sends telemetry back to cloud", "ComponentPublisher": "[Your Company]", "ComponentConfiguration": { "DefaultConfiguration": { "accessControl": { "aws.greengrass.ipc.mqttproxy": { "com.example.DeviceController:mqttproxy:1": { "policyDescription": "Allows access to subscribe to command topics from IoT Core", "operations": [ "aws.greengrass#SubscribeToIoTCore" ], "resources": [ "commands/device1" ] }, "com.example.DeviceController:mqttproxy:2": { "policyDescription": "Allows access to publish telemetry to IoT Core", "operations": [ "aws.greengrass#PublishToIoTCore" ], "resources": [ "telemetry/*" ] } } } } }, "Manifests": [ { "Platform": { "os": "linux", "runtime": "*" }, "Lifecycle": { "run": "cd {artifacts:decompressedPath}/DeviceController && node DeviceController.js" }, "Artifacts": [ { "Uri": "s3://YOUR-BUCKET/artifacts/com.example.DeviceController/1.0.0/DeviceController.zip", "Unarchive": "ZIP" } ] } ] }
C/C++
{ "RecipeFormatVersion": "2020-01-25", "ComponentName": "com.example.DeviceController", "ComponentVersion": "1.0.0", "ComponentType": "aws.greengrass.generic", "ComponentDescription": "Receives commands from IoT Core and sends telemetry back to cloud", "ComponentPublisher": "[Your Company]", "ComponentConfiguration": { "DefaultConfiguration": { "accessControl": { "aws.greengrass.ipc.mqttproxy": { "com.example.DeviceController:mqttproxy:1": { "policyDescription": "Allows access to subscribe to IoT Core topics", "operations": ["aws.greengrass#SubscribeToIoTCore"], "resources": ["commands/device1"] }, "com.example.DeviceController:mqttproxy:2": { "policyDescription": "Allows access to publish to IoT Core topics", "operations": ["aws.greengrass#PublishToIoTCore"], "resources": ["telemetry/device1"] } } } } }, "Manifests": [ { "Platform": { "os": "linux", "runtime": "*" }, "Lifecycle": { "run": "{artifacts:path}/device_controller" }, "Artifacts": [ { "Uri": "s3://YOUR-BUCKET/artifacts/com.example.DeviceController/1.0.0/device_controller" } ] } ] }