AWS Certified Data Analytics – Specialty — Question 86

A company wants to collect and process events data from different departments in near-real time. Before storing the data in Amazon S3, the company needs to clean the data by standardizing the format of the address and timestamp columns. The data varies in size based on the overall load at each particular point in time. A single data record can be 100 KB-10 MB.
How should a data analytics specialist design the solution for data ingestion?

Answer options

• A. Use Amazon Kinesis Data Streams. Configure a stream for the raw data. Use a Kinesis Agent to write data to the stream. Create an Amazon Kinesis Data Analytics application that reads data from the raw stream, cleanses it, and stores the output to Amazon S3.
• B. Use Amazon Kinesis Data Firehose. Configure a Firehose delivery stream with a preprocessing AWS Lambda function for data cleansing. Use a Kinesis Agent to write data to the delivery stream. Configure Kinesis Data Firehose to deliver the data to Amazon S3.
• C. Use Amazon Managed Streaming for Apache Kafka. Configure a topic for the raw data. Use a Kafka producer to write data to the topic. Create an application on Amazon EC2 that reads data from the topic by using the Apache Kafka consumer API, cleanses the data, and writes to Amazon S3.
• D. Use Amazon Simple Queue Service (Amazon SQS). Configure an AWS Lambda function to read events from the SQS queue and upload the events to Amazon S3.

Correct answer: C

Explanation

The correct answer is C because using Amazon Managed Streaming for Apache Kafka allows for efficient handling of varying data sizes and provides a robust way to cleanse and process data before storing it in Amazon S3. Options A and B are more suited for real-time analytics rather than batch processing, while option D does not provide the necessary data cleansing capabilities at scale.