DEA-C01 Question 279
Select 3You are designing a data lake architecture for a media company that processes multiple types of data, including structured metadata from relational databases, semi-structured JSON files from web applications, and unstructured video files. Which of the following data modeling approaches should you use to effectively organize and process these data types in AWS?
- A
Store structured data in Amazon RDS or Amazon Redshift for analytics and reporting.
- B
Store semi-structured JSON files in Amazon S3 and query them using Amazon Athena or AWS Glue.
- C
Store unstructured video files in Amazon Elasticsearch Service for full-text search capabilities.
- D
Use Amazon S3 to store unstructured video files and leverage Amazon Rekognition for video analysis.
- E
Convert all data types into a single format, such as Parquet, to ensure consistency.
Show answer and explanation
Correct answers: A, B, D
Explanation
Data modeling for structured, semi-structured, and unstructured data requires selecting appropriate AWS services that cater to the unique characteristics of each data type. Structured data fits well into relational databases like Amazon RDS or Redshift. Semi-structured data formats, such as JSON, are well-suited for storage in Amazon S3 and querying with tools like Athena or Glue. Unstructured data, such as video files, can be effectively stored in Amazon S3, with additional analysis capabilities provided by AI services like Amazon Rekognition. This multi-service approach ensures optimal performance and cost-efficiency.
- A. Correct.
Structured data, such as relational database metadata, is best stored in Amazon RDS or Amazon Redshift for efficient querying and analytics.
- B. Correct.
Semi-structured data like JSON files can be stored in Amazon S3 and queried using serverless tools like Amazon Athena or AWS Glue for ETL operations.
- C. Incorrect.
Amazon Elasticsearch Service is designed for search and analytics of text-based data, not for storing and processing unstructured video files.
- D. Correct.
Unstructured video files can be stored cost-effectively in Amazon S3, and Amazon Rekognition can be used for AI-based video analysis.
- E. Incorrect.
Converting all data types into a single format, such as Parquet, is not practical or efficient because different data types have unique storage and processing requirements.