You work for a farming equipment component manufacturer which builds farm product containers like corn silos, milk containers, etc. These containers have IoT sensors built into them that transmit information such as fill rate, capacity usage, etc. The IoT devices transmit their data back to your cloud environment via the MQTT protocol. You want to use a machine learning model to predict container usage by region and by product stored. This information will help your management team use real-time dashboards to better understand their product marketing campaigns by region. You have created, trained, and deployed to Amazon SageMaker Hosting Services your model based on the Linear Learner algorithm.
What AWS services would you use to create your pipeline to feed your inference requests to your model? (Select THREE)

Answer options:

A.IoT Greengrass to receive the IoT device MQTT messages
B.IoT Core to receive the IoT device MQTT messages
C.Elastic Beanstalk to stream the IoT messages
D.Kinesis Data Streams to stream the IoT messages
E.A Lambda function to transform the IoT message data to the inference request serialization format
F.API Gateway to transform the IoT message data to the inference request serialization format
G.Route 53 to stream the IoT messages

Answers: B, D, E
Option A is incorrect. AWS IoT Greengrass is used to extend AWS to edge devices, such as your sensors in your farming containers. Greengrass is used to perform prediction directly on the devices themselves. IoT Core is a better option for receiving your MQTT IoT messages for processing via your machine learning inference running in Amazon SageMaker Hosting Services.
Option B is correct. IoT Core is designed to allow IoT devices to interact with other AWS services, such as Kinesis Data Streams.
Option C is incorrect. Elastic Beanstalk is used to host web applications or worker nodes for web applications. You wouldn’t use Elastic Beanstalk to stream IoT messages.
Option D is correct. Kinesis Data Streams can receive IoT messages from IoT Core and stream them to your SageMaker inference endpoint via a Lambda function.
Option E is correct. You can write a lambda function that is triggered by Kinesis Data Streams that transforms your IoT messages into the inference serialization format required by your inference endpoint.
Option F is incorrect. API Gateway is used to create an API request endpoint. You wouldn’t use API Gateway to transform IoT message data. You would have to have a lambda function behind your API Gateway to accomplish this.
Option G is incorrect. Route 53 is Amazon’s DNS server implementation. You can’t use Route 53 to transform messages.
Reference:
Please see the Amazon SageMaker developer guide titled Train a Model with Amazon SageMaker, the AWS IoT Core overview, the AWS IoT developer guide titled Creating an AWS Lambda Rule, and the AWS IoT Greengrass overview.
Here is a diagram of the proposed solution: