A media company needs to process large volumes of video metadata (XML files) generated hourly from its content management system. These files are stored in an on-premises SFTP server. The company wants to automate the ingestion of these files into Amazon S3, parse the XML, extract specific metadata fields, and store the extracted data in a structured format (Parquet) in a separate S3 bucket for analytics. The process should be fully managed and serverless. Which combination of AWS services should be used?

Why the other options are wrong

• A. AWS Direct Connect provides a dedicated network connection, but not an SFTP endpoint. Amazon EMR offers powerful processing (e.g., Spark), but it requires managing clusters, which goes against the serverless requirement of the transformation step.
• C. AWS Storage Gateway (File Gateway) provides file-based access to S3 but doesn't offer a direct SFTP endpoint for external systems. Amazon Redshift is a data warehouse, not a transformation engine for files in S3.
• D. AWS DataSync is suitable for large-scale migrations and ongoing replication, but AWS Transfer Family is specifically designed for SFTP access, which is the exact source system. AWS Glue ETL is appropriate for transformation.