Databricks-DE-Pro Cost and Performance Optimization 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.

• ✗

  Load all data into a high-performance in-memory database.

  Why it's wrong here

  In-memory databases are extremely expensive due to the high cost of RAM. For rarely updated data, this is an unnecessary expense, as the performance gains of in-memory processing are not required for static or cold datasets. It is a highly inefficient use of budget for analytical storage.

• ✓

  Store the data in Delta format on cloud object storage.

  Why this is correct

  Object storage is highly durable and inexpensive. Delta Lake allows you to query this data with high performance using Databricks SQL while keeping the storage costs at the lowest possible tier. This is the standard, most cost-effective architecture for large, rarely updated datasets in a modern data lakehouse.

• ✗

  Replicate the data into multiple cloud regions for higher availability.

  Why it's wrong here

  Cross-region replication significantly increases storage costs and adds complexity to synchronization. Unless explicitly required for disaster recovery or global latency needs, it is an unnecessary expense for rarely updated data that can be managed within a single region, wasting budget that could be used for other purposes.

• ✗

  Convert the table to a legacy Hive format for better compatibility.

  Why it's wrong here

  Legacy Hive formats lack the advanced performance optimizations of Delta Lake, such as data skipping, Z-Ordering, and efficient schema enforcement. Using an outdated format will result in slower queries, requiring more compute resources and higher costs to achieve the same analytical performance as a modern Delta table.