2 3 4 5 6 7 8 Sensors collect data to send to the device running AWS IoT Greengrass. AWS IoT Greengrass and AWS Snowball Edge devices handle intermittent connectivity and low latency inference. Sensor data is collected by a AWS Lambda producer function. You can use container images from public, private, or Amazon Elastic Container Registry (Amazon ECR), or Docker Hub. Amazon SageMaker Neo enables machine learning models to execute on the device with a compact runtime. Action model inference uses AWS Lambda for control plane enablement. Deploy AWS IoT Greengrass connectors to use Amazon Simple Notification Service (Amazon SNS). AWS IoT Greengrass supports Neooptimized models so that you can deploy your models directly to the edge with over-the-air update
