AWS Certified Solutions Architect – Associate (SAA-C02) — Question 387
A company is developing a new machine learning model solution in AWS. The models are developed as independent microservices that fetch about 1 GB of model data from Amazon S3 at startup and load the data into memory. Users access the models through an asynchronous API. Users can send a request or a batch of requests and specify where the results should be sent.
The company provides models to hundreds of users. The usage patterns for the models are irregular Some models could be unused for days or weeks. Other models could receive batches of thousands of requests at a time.
Which solution meets these requirements?
Answer options
- A. The requests from the API are sent to an Application Load Balancer (ALB). Models are deployed as AWS Lambda functions invoked by the ALB.
- B. The requests from the API are sent to the models Amazon Simple Queue Service (Amazon SQS) queue. Models are deployed as AWS Lambda functions triggered by SQS events AWS Auto Scaling is enabled on Lambda to increase the number of vCPUs based on the SQS queue size.
- C. The requests from the API are sent to the model's Amazon Simple Queue Service (Amazon SQS) queue. Models are deployed as Amazon Elastic Container Service (Amazon ECS) services reading from the queue AWS App Mesh scales the instances of the ECS cluster based on the SQS queue size.
- D. The requests from the API are sent to the models Amazon Simple Queue Service (Amazon SQS) queue. Models are deployed as Amazon Elastic Container Service (Amazon ECS) services reading from the queue AWS Auto Scaling is enabled on Amazon ECS for both the cluster and copies of the service based on the queue size.
Correct answer: D
Explanation
AWS Lambda is not suitable for this scenario because downloading and loading 1 GB of model data during startup would cause unacceptable latency during cold starts, ruling out options A and B. Amazon ECS is the correct environment for these heavyweight containers, and Amazon SQS is ideal for decoupling and handling the asynchronous, irregular batch requests. Option D is correct because AWS Auto Scaling can scale both the ECS tasks and the underlying container instances based on the SQS queue depth, whereas App Mesh in option C is a service mesh and cannot scale ECS cluster instances.