Exam AWS Certified Machine Learning - Specialty topic 1 question 71 discussion

B it is .

A is not correct because Minimize the number of servers that will need to be managed. EMR is not server-less. B is correct. AWS Glue supports an extension of the PySpark Python dialect for scripting extract, transform, and load... C is not correct because using Lambda for ETL you will not be able to Reuse existing PySpark logic D is not correct because Kinesis is not server-less. And you can not Reuse existing PySpark logic

Option B (using AWS Glue for the ETL process) is the best solution for the described requirements. A: This solution requires managing an Amazon EMR cluster, which would involve more server management than AWS Glue, violating the requirement to minimize the number of servers to be managed. C: AWS Lambda is not ideal for this use case because it has resource limitations, including memory and execution time limits (15 minutes max), which might not be suitable for large-scale ETL operations involving PySpark logic. D: Amazon Kinesis Data Analytics is focused on real-time stream processing, which doesn't fit the described scheduled batch processing scenario.

Answer is A, as B clearly mentions that Pyspark code is written with leverage from already existing code. Also, the server architecture used currently is on-premises which will have more servers that solution A.

Amazon Kinesis Data Analytics is more suited for real-time processing and streaming data. The given use case does not indicate a need for real-time processing, so this might not be the best fit. Furthermore, it doesn't support PySpark natively.

Voted B based on the serverless (minimum servers) and https://docs.aws.amazon.com/glue/latest/dg/aws-glue-programming.html

Indeed B using Glue

B is the correct. A you have to manage EMR, so it's wrong. D you don't use Spark, so it's wrong. C you will not be using Spark, so it's wrong.

B ticks all boxes. Minimize servers -> AWS managed services -> Glue.

Option A would be the best response for this scenario. This solution allows the Data Scientist to reuse the existing PySpark logic while migrating the ETL process to the cloud. The raw data is written to Amazon S3, and a Lambda function is scheduled to trigger a Spark step on a persistent EMR cluster based on the existing schedule. The PySpark logic is used to run the ETL job on the EMR cluster, and the results are output to a processed location in Amazon S3 that is accessible for downstream use. This solution minimizes the number of servers that need to be managed, and it allows for a seamless migration of the existing ETL process to the cloud.

Option D is wrong it should be B

D cannot be answer as there is no streaming data or Realtime processing.

the answer is b

Answer should be B. Serverless, on a regular schedule (no real time requirement), reuses PySpark code in Glue ETL script.

Answer is B as they specifically ask about reusing existing PySpark, which can be done with Glue

https://docs.aws.amazon.com/glue/latest/dg/creating_running_workflows.html

It is B. ! "Minimize number of servers to be managed". B is a Serverless solution which fulfils other requirements!