Which of the following zones within a data lake requires sensitive data to be encrypted or tokenized?

The raw zone of a data lake is the ingestion layer where data lands in its original, unaltered format from various source systems. Because this data has not yet been processed, cleansed, or classified, it poses the highest risk of containing sensitive or personal information. To adhere to security and privacy principles like "secure by design," sensitive data must be protected at the earliest possible point. Therefore, applying encryption or tokenization in the raw zone is a critical control to prevent unauthorized access or data spillage before it propagates through the lake.

A. The trusted zone contains cleansed and curated data; protection should have been applied before data reached this stage.
B. The clean zone is similar to the trusted zone, where data has already undergone processing and should already be protected.
D. A temporal zone is a staging area; while security is important, the initial and most critical protection point is the raw zone.

1. ISACA, CDPSE Review Manual, 2020, Chapter 2, Section: Data Lifecycle, "Protection of data should be applied throughout the data life cycle, from creation to destruction. This includes applying controls at the point of ingestion into systems like data lakes."

2. AWS Whitepaper, Build a modern data architecture with AWS Lake House, 2021, p. 6, "The raw zone stores data in its original format...Data is encrypted and partitioned in the raw zone."

3. Microsoft Azure Documentation, Data lake zones and containers, 2023, "Raw data (or landing zone) - Data is ingested in its original, raw format...Data should be immutable and can be stored in its original format, such as JSON or .csv." This implies protection is paramount as the state is unknown.