A network security vendor needs to ingest telemetry data from millions of endpoints running all over the world. This data is transmitted every 30 seconds in the form of records containing 50 fields…

Question

Options

• AConfigure Kinesis Data Streams to send the data through a stream to AWS Lambda.
• BConfigure Kinesis Data Streams to send the data through a stream to aggregate the data hourly
• CConfigure Kinesis Data Streams to send the data through a stream to aggregate the data hourly
• DConfigure Kinesis Data Streams to send the data through a stream to aggregate the data hourly

• A
  3% (1)
• B
  9% (3)
• C
  72% (23)
• D
  16% (5)

Explanation

The requirements are: aggregate records hourly and store in a format optimized for Athena queries targeting 7-12 out of 50 columns. Athena queries columnar formats (Apache Parquet or ORC) far more efficiently than row-based formats, because only the queried columns are read from storage, dramatically reducing scan costs and time. Kinesis Data Firehose natively supports dynamic partitioning and data format conversion to Parquet/ORC using AWS Glue, requiring minimal custom code. This makes option C (hourly aggregation via Firehose with columnar output) the approach with the least customization. Option A (Lambda) requires custom transformation code to be written and maintained. Option B (aggregating without columnar conversion) would not optimize Athena's column-selective queries. Option D implies a different implementation that requires more custom work. The built-in Firehose-to-Parquet conversion via Glue is the lowest-effort, purpose-built solution.

No community discussion yet for this question.