You are building a real-time prediction engine that streams files which may contain Personally Identifiable Information (PII) to Google Cloud. You want to use the Cloud Data Loss Prevention (DLP)…

Question

Options

• AStream all files to Google Cloud, and then write the data to BigQuery.
• BStream all files to Google Cloud, and write batches of the data to BigQuery.
• CCreate two buckets of data: Sensitive and Non-sensitive.
• DCreate three buckets of data: Quarantine, Sensitive, and Non-sensitive.

• A
  9% (3)
• B
  3% (1)
• C
  6% (2)
• D
  82% (27)

Explanation

The three-bucket architecture (Quarantine, Sensitive, Non-sensitive) is the recommended pattern for DLP-integrated pipelines. Incoming files are first written to the Quarantine bucket - a staging area where no unauthorized access is permitted. The Cloud DLP API then scans each file in the Quarantine bucket. Based on the scan results, files are routed to either the Sensitive bucket (confirmed PII found, access restricted to authorized personnel) or the Non-sensitive bucket (no PII detected, broader access permitted). This ensures that PII is never exposed to unauthorized users even momentarily, because all data passes through the controlled Quarantine stage before being classified. Options A and B write directly to BigQuery without a DLP scan gate, meaning PII could be exposed before it is identified. Option C uses only two buckets and lacks the Quarantine staging layer that prevents premature access.

No community discussion yet for this question.