SnowPro Advanced: Security Engineer Question 322
Select 2Implement mitigation strategiesA security engineer discovers that several analysts have been running ad hoc queries directly against a sensitive customer table that contains columns with PII. The analysts need continued access to most of the dataset for reporting, but the company must reduce the risk of unnecessary exposure while minimizing application changes. Which TWO actions provide the most effective mitigation strategy in Snowflake?
- A
Apply a masking policy to the PII columns and make the policy role-aware so only approved roles can see unmasked values.
- B
Create a secure view that excludes or transforms the sensitive columns, and grant analysts access to the secure view instead of the base table.
- C
Move the table to a different virtual warehouse so only authorized users can query the sensitive rows.
- D
Increase the network policy restrictions for analyst users so they can connect only from corporate IP addresses.
- E
Grant analysts OWNERSHIP on the table so they can manage access without opening a support ticket.
Show answer and explanation
Correct answers: A, B
Explanation
The best mitigation strategy is to reduce exposure at the data access layer while preserving needed business access. In Snowflake, masking policies are the standard control for protecting sensitive column values based on context such as the active role. Secure views are also a strong mitigation because they let teams expose only the required subset of data and can be granted instead of the base table. Together, these controls align with least privilege and data minimization principles.
By contrast, warehouses are for compute isolation and performance management, not authorization. Network policies are valuable for reducing attack surface by restricting connection locations, but they do not prevent an already authorized user from seeing sensitive columns. Granting OWNERSHIP is excessive and violates security best practices.
Relevant Snowflake best practices and documentation areas include Dynamic Data Masking, Secure Views, Role-Based Access Control, and Network Policies. For this scenario, the mitigation should focus on column-level and object-level exposure reduction rather than compute placement or broader administrative privileges.
- A. Correct.
Correct. Dynamic data masking is a primary Snowflake mitigation control for limiting exposure of sensitive column values without requiring major downstream changes. A role-aware masking policy can reveal cleartext only to approved roles and return masked or tokenized values to other users. This directly addresses the requirement to let analysts continue working with the dataset while reducing unnecessary PII exposure.
- B. Correct.
Correct. A secure view is an effective mitigation when analysts do not need direct access to the base table. The secure view can omit sensitive columns or present transformed values, and secure views are designed to prevent exposing underlying query logic or data through certain optimizations. Granting access to the secure view instead of the table follows least-privilege principles and minimizes application changes when reports can be redirected to the view.
- C. Incorrect.
Incorrect. Virtual warehouses control compute resources, not data authorization. Moving a table to a different warehouse does not restrict who can see particular tables, rows, or columns. This option reflects a common misconception that warehouse isolation provides data security controls; in Snowflake, access is governed through roles, grants, and data protection features such as masking policies and views.
- D. Incorrect.
Incorrect. Network policies can reduce access risk by restricting where users connect from, and they are useful as a defense-in-depth measure. However, they do not solve the core problem described: authorized analysts are still seeing sensitive PII they do not need. Network restrictions mitigate connection risk, not overexposure of columns within queries.
- E. Incorrect.
Incorrect. Granting OWNERSHIP would expand privileges significantly and increase risk. OWNERSHIP is the highest level of control over an object and is contrary to least-privilege design. It would not mitigate unnecessary PII exposure; it would make the exposure problem worse by allowing analysts to alter grants and object definitions.