1Z0-184-25 Question 70
Single answerYou are building a recommendation engine for an e-commerce platform in Oracle Cloud Infrastructure. Your image-processing pipeline generates embedding vectors for millions of product images stored in Object Storage, and you need a way to quickly retrieve visually similar products based on these embeddings. Which solution is recommended for performing high-scale, low-latency similarity searches on these vectors in OCI?
- A
Store embeddings as files in Object Storage and run custom similarity search code on Oracle Functions
- B
Use Oracle Identity and Access Management (IAM) to index and query vectors directly
- C
Use Oracle Logging Analytics to process embeddings and perform query-based similarity search
- D
Use Oracle Autonomous Database with built-in vector search capability to store and query embedding vectors
Show answer and explanation
Correct answer: D
Explanation
Oracle Autonomous Database (particularly with the newer vector search functionality introduced in recent database releases) is designed for high-volume and low-latency vector similarity queries. This approach simplifies the process of storing embeddings, creating indexes, and running real-time similarity searches. For more details, consult the official Oracle documentation on Autonomous Database support for vector search features and best practices for indexing large sets of embeddings.
- A. Incorrect.
Option 1: Storing embeddings as files in Object Storage does not provide an efficient built-in index or query engine for vectors. While you can run custom code on Oracle Functions, it would require handling all indexing logic in your function� memory, which is not scalable or performant for large data sets.
- B. Incorrect.
Option 2: IAM is specifically designed for managing identities, policies, and access controls. It is not intended for data storage or conducting vector-based similarity searches. Therefore, this option is not correct for embedding queries.
- C. Incorrect.
Option 3: Oracle Logging Analytics enables analysis of log data and machine-generated data, not vector embeddings for similarity search. It does not provide the specialized indexing needed for real-time content-based queries.
- D. Correct.
Option 4: Oracle Autonomous Database supports vector data storage and similarity search capabilities. By creating appropriate vector indexes and leveraging the built-in functionality, you gain low-latency, high-scale similarity queries on embedding vectors, making this the most suitable choice.