Exam AWS Certified Machine Learning - Specialty topic 1 question 61 discussion

I WOULD LIKE TO CHOOSE ANSWER A. https://aws.amazon.com/tw/blogs/machine-learning/use-the-built-in-amazon-sagemaker-random-cut-forest-algorithm-for-anomaly-detection/

Answer is A. As the word anamoly talks about Random Cut Forest in the exam and that can be done in a cost effective manner using Kinesis Data Analytics

The question says REAL TIME events doesn't that eliminate Data Firehose as it is technically NEAR real time but not real time like Data Stream? Though Random Cut Forest seems like the best option for anomaly detection. I'm torn between A and B

Kinesis Firehose and Data Analytics with random cut forest should do it.

A. Based on these considerations, Option A is the most efficient way to accomplish the tasks. It provides a seamless, real-time data ingestion and processing pipeline, leverages machine learning for anomaly detection, and efficiently stores data in a data lake, meeting all the key requirements of the cybersecurity company.

ONLY A

B not as efficient for real-time processing and storing results as using Kinesis services.

At least B is a possible solution, but A will not work as KDF doesn't support KDA as a destination service https://docs.aws.amazon.com/firehose/latest/dev/create-name.html . In my opinion, KDF should always be the latest Kinesis Service in a streaming pipeline

A has all the required steps

A. YES - Firehose can pipe into KDA, and KDA supports RCF B. NO - RCF best for anomality detection C. NO - no need for intermediary S3 storage D. NO - no need for intermediary S3 storage

option A

A is the correct. One tip for the exam: When you see Data Streaming, possibly the solution should contains a Kinesis Service. B is too much complex!

Makes sense to select A here.

I strongly believe A is the right answer. At a minimum there should be some justification provided for your answer.

Amazon Kinesis Data Firehose is a fully managed service for streaming real-time data to Amazon S3 and can handle the ingestion of large amounts of data in real time. Kinesis Data Analytics Random Cut Forest (RCF) is a fully managed service that can be used to perform anomaly detection on streaming data, making it well suited for this use case. The results of the anomaly detection can then be streamed to Amazon S3 using Kinesis Data Firehose, providing a scalable and cost-effective data lake for later processing and analysis.

I would select A

B is too resource intensive for that use case. I choose A, but I think the data should be better ingested using Kinesis streams