ARA-C01 Question 129
Single answer2.2 Design data sharing solutions, based on different use cases.A healthcare analytics company stores PHI and non-PHI data in a Snowflake account on AWS us-east-1. It must provide near real-time access to a curated, non-PHI dataset to three external organizations. Two partners already use Snowflake in different regions and cloud platforms, while the third partner does not have a Snowflake account. The provider wants to minimize ongoing data movement, avoid creating separate ETL pipelines for each partner, and ensure that each partner can query the same governed dataset independently. Which solution best meets these requirements?
- A
Create a secure data share from the provider account, and use listings/reader account capabilities as needed so Snowflake consumers can access the shared database without copying data; use cross-region/cloud data sharing where supported for the existing Snowflake partners and a reader account for the partner without Snowflake.
- B
Unload the curated dataset to cloud object storage for each partner, then configure each partner to ingest the files into their own analytics platform on a recurring schedule.
- C
Replicate the entire source database into separate Snowflake accounts owned by the provider for each partner, and grant each partner direct access to the replicated databases.
- D
Create one shared database inside the provider account and give all three external organizations user accounts in that same provider account so they can query the governed data directly.
Show answer and explanation
Correct answer: A
Explanation
The best solution is Snowflake Secure Data Sharing, extended with the appropriate consumer-access model. Secure shares allow a provider to expose live, governed data to other Snowflake accounts without copying the underlying data. Consumers create a database from the share and query it using their own warehouses, which preserves independent performance management and access control on the consumer side. For non-Snowflake customers, reader accounts let the provider share data without requiring the consumer to purchase or administer a Snowflake account. For accounts in different regions or cloud platforms, Snowflake supports broader sharing patterns through cross-region/cross-cloud capabilities and listings, avoiding bespoke ETL pipelines. This design is aligned with Snowflake best practices for data sharing: use shares for live governed access, use reader accounts for consumers without Snowflake, and avoid file exports or per-consumer duplicated pipelines unless there is a specific non-Snowflake integration requirement. Relevant Snowflake documentation includes Secure Data Sharing, Reader Accounts, and Sharing Data Across Regions and Cloud Platforms.
- A. Correct.
Correct. Secure Data Sharing is designed to let providers share live data without copying or moving the underlying tables for each consumer. Consumers query the shared objects independently using their own compute, which satisfies the requirement that each partner query the same governed dataset independently. For organizations already on Snowflake, the provider can use direct shares and, when consumers are in different regions/clouds, Snowflake's cross-cloud/cross-region sharing mechanisms via listings/replication-supported sharing patterns. For the organization without a Snowflake account, a reader account is the appropriate option because it allows data access without requiring the consumer to be a Snowflake customer. This approach minimizes ETL and ongoing data movement compared with file-based distribution.
- B. Incorrect.
Incorrect. Unloading to object storage and having each partner ingest the data creates separate distribution pipelines, increases operational overhead, introduces data latency, and results in copied data. That directly conflicts with the stated goals of minimizing ongoing data movement and avoiding separate ETL pipelines per partner. While file exchange is possible, it is not the best Snowflake-native data sharing design for this use case.
- C. Incorrect.
Incorrect. Replicating the entire source database into separate provider-owned accounts for each partner is heavier than necessary and does not align with the requirement to avoid unnecessary data movement. Database replication is useful for disaster recovery, business continuity, or certain cross-region availability patterns, but it is not the primary mechanism for giving multiple external consumers governed access to the same curated dataset. It also increases administration and cost compared with secure sharing.
- D. Incorrect.
Incorrect. Giving external organizations users in the provider's account is generally a poor architecture for B2B data sharing. It weakens tenant isolation, complicates governance and billing boundaries, and does not provide each consumer with an independent Snowflake consumption model. Snowflake best practice for external sharing is to use Secure Data Sharing, listings, or reader accounts rather than co-mingling external consumers as users inside the provider account.