Exam AWS Certified Machine Learning - Specialty topic 1 question 270 discussion

Easiest option would be C

A. NO - Kinesis Data Analytics can use only Firehose or Amazon Kinesis Data Streams as input (https://docs.aws.amazon.com/kinesisanalytics/latest/dev/how-it-works.html) B. NO - no need to save in S3, can do on-the-fly C. YES D. NO - AWS Lambda would be invoked on a per-record basis

why not B?

Option C also allows the ML specialist to use Amazon Kinesis Data Analytics to transform the most recent 10 minutes of data before inference. Kinesis Data Analytics is a fully managed service that enables users to analyze streaming data using SQL or Apache Flink. Kinesis Data Analytics can process streaming data in real time and generate insights, metrics, and alerts. Kinesis Data Analytics can also integrate with other AWS services, such as Lambda, S3, or SageMaker. The ML specialist can use Kinesis Data Analytics to apply SQL queries or Flink applications to transform the event data based on the 10-minute running window and prepare it for inference.

It’s real-time and less operational overhead

I think it's C as we are using only 2 services and it's less operational effort.