ARA-C01 Snowflake Architecture Practice Question

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

• ✗

  Use Snowflake's multi-cluster warehouse feature to deploy warehouses in each region that read from the primary database.

  Why it's wrong here

  Multi-cluster warehouses scale compute within a single region; they do not deploy warehouses in multiple regions. Warehouses in other regions cannot directly access the primary database without replication. This approach would still incur cross-region data transfer and latency.

• ✗

  Configure a Snowflake materialized view in each region that refreshes from the primary table.

  Why it's wrong here

  Materialized views are stored within the same database and region; they cannot be created in a different region from the base table. They also consume storage and compute for maintenance. They do not provide cross-region replication.

• ✗

  Enable Snowflake's global data sharing by creating a share and granting access to accounts in each region.

  Why it's wrong here

  Data sharing allows other accounts to read data, but the data remains in the provider's region. Consumers in other regions would still access data cross-region, incurring latency and data transfer costs. It does not create a local copy.

• ✓

  Create a read-only replica of the primary database in each region using database replication and configure clients to connect to the nearest replica.

  Why this is correct

  Database replication creates a read-only copy of the database in another region. Clients in Europe and Asia can query the local replica, reducing latency and avoiding cross-region data transfer costs for read queries. This is the recommended approach for global read-heavy workloads with low latency requirements.