A company has decided to move some workloads onto AWS to create a grid environment to run market analytics. The grid will consist of many similar instances, spun-up by a job-scheduling function. Each time a large analytics workload is completed, a new VPC is deployed along with job scheduler and grid nodes. Multiple grids could be running in parallel.
Key requirements are:
✑ Grid instances must communicate with Amazon S3 retrieve data to be processed.
✑ Grid instances must communicate with Amazon DynamoDB to track intermediate data,
✑ The job scheduler need only to communicate with the Amazon EC2 API to start new grid nodes.
A key requirement is that the environment has no access to the internet, either directly or via the on-premises proxy. However, the application needs to be able to seamlessly communicate to Amazon S3, Amazon DynamoDB, and Amazon EC2 API, without the need for reconfiguration for each new deployment.
Which of the following should the Solutions Architect do to achieve this target architecture? (Choose three.)

• A. Enable VPC endpoints for Amazon S3 and DynamoDB.
• B. Disable Private DNS Name Support.
• C. Configure the application on the grid instances to use the private DNS name of the Amazon S3 endpoint.
• D. Populate the on-premises DNS server with the private IP addresses of the EC2 endpoint.
• E. Enable an interface VPC endpoint for EC2.
• F. Configure Amazon S3 endpoint policy to permit access only from the grid nodes.

感觉 AE 是对的，CF 之间怎么取舍 ？都有道理。。