1Z0-184-25 Question 94
Single answerYou are developing a recommendation system on Oracle Cloud Infrastructure (OCI) that relies on vector embeddings to determine similarity between content items. Your team wants to perform efficient, low-latency similarity searches against a rapidly growing dataset of these embeddings. Which solution provides the best approach for storing and querying vector embeddings in this scenario?
- A
Store the embeddings in Oracle Database 23c with vector search indexes enabled
- B
Write the embeddings as individual files to Object Storage for quick ad-hoc retrieval
- C
Use HPC compute shapes to keep embeddings in local memory for faster lookups
- D
Leverage MySQL Database Service without specialized indexing for advanced queries
Show answer and explanation
Correct answer: A
Explanation
Oracle Database 23c introduces native vector search capabilities for fast similarity lookups on vector data. This feature allows you to store embeddings directly in the database and leverage specialized indexing for high-performance retrieval. For more details, refer to the Oracle Database 23c documentation on Vector Search, which covers best practices for storing, indexing, and querying vector embeddings in real-world applications.
- A. Correct.
Correct. Oracle Database 23c includes built-in vector search indexes that are specifically designed to efficiently store and query vector embeddings. This approach provides high performance for similarity searches and integrates well with the rest of your OCI workflow.
- B. Incorrect.
Incorrect. Storing vector embeddings as files in Object Storage is feasible for archiving or backup, but it is not optimized for real-time similarity queries, where you need fast indexing and retrieval. Object Storage lacks the specialized indexing structures needed for low-latency searches.
- C. Incorrect.
Incorrect. While HPC compute shapes provide high performance computing resources, relying on volatile local memory to store embeddings is not practical for a growing dataset that requires persistent storage. If you lose the instance, you lose all embeddings in local memory as well.
- D. Incorrect.
Incorrect. MySQL Database Service supports traditional relational queries but does not offer built-in vector indexing features at this time. Storing embeddings here would require manual or third-party indexing solutions, which may not be as performant and straightforward as using Oracle Database 23c� native vector search capabilities.