关于Greengrass组仅含多Core设备及Lambda订阅配置的技术问询

Hey there! Let's break down your two questions about AWS Greengrass multi-core groups and Lambda subscriptions:

Absolutely! AWS Greengrass fully supports multi-core Greengrass groups where all devices in the group are Greengrass Core devices—you don’t need to include any edge (non-core) devices at all.

Here’s what you need to keep in mind:

• Each core device must first be registered in your AWS account, with the Greengrass Core software installed and properly configured.
• When setting up the group, during the device addition step, just select "Greengrass core" for every device you want to include.
• Multi-core groups are specifically designed for use cases like distributed processing, cross-core device communication, or running redundant workloads across multiple core devices.

Since you need Lambda functions on different cores to send and receive messages, let’s cover both intra-core (same device) and inter-core (different devices) communication scenarios:

Intra-core Lambda communication (same core device) #

If you want instances of the same Lambda on the same core to communicate with each other:

• Create a subscription with these settings:
  • Source: The ARN (or function name) of your Lambda.
  • Target: The exact same Lambda ARN/function name.
  • Topic: A custom MQTT topic (e.g., local/lambda/intra-core/messages).
• When the Lambda publishes to this topic, any running instance of the same Lambda on that core will receive the message.

Inter-core Lambda communication (across different core devices) #

For Lambda functions on different cores to exchange messages, you’ll leverage Greengrass’s cross-core message routing capabilities:

1. First, deploy the Lambda to all core devices:
   • In your Greengrass group’s Lambda deployment configuration, select all core devices as the target for the Lambda. Use a consistent version or alias to ensure all cores run identical code.
2. Configure subscriptions for cross-core messaging:
   • Option 1: Point-to-point communication
     • Set Source to your Lambda’s ARN/function name.
     • Set Target to the same Lambda ARN/function name.
     • Use a topic that includes core identifiers (e.g., core/{core-device-id}/lambda/messages) if you want to send messages to a specific core’s Lambda instance.
   • Option 2: Broadcast to all cores’ Lambda instances
     • Use a broad topic like core/all/lambda/broadcast.
     • Configure a subscription where Source is your Lambda, Target is your Lambda, and Topic is this broadcast topic. All cores’ Lambda instances that subscribe to this topic will receive the message.
3. Key reminders:
   • Ensure your Lambda has the necessary IAM permissions to publish and subscribe to the MQTT topics you define.
   • When deploying the group configuration, make sure the subscription changes are included in the deployment to all core devices.

内容的提问来源于stack exchange，提问作者techy