Difference between Azure AI Studio Preview Indexes and Search Service Indexes?

Question

In Azure AI Studio Preview I have some indexes:

And also in my search service:

What is the difference between these two types of indexes?

When creating one Azure AI Studio Preview it takes roughly 20 minutes to 1 hour to create it. When creating one in Search Service it's a matter of less than 30 seconds. Why such a big difference?

(It's also not clear how to delete an index from Azure AI Studio Preview)

Answer 1

Hi Dirk Broenink,

Thanks for reaching out to Microsoft Q&A.

What is the difference between these two types of indexes? When creating one Azure AI Studio Preview it takes roughly 20 minutes to 1 hour to create it. When creating one in Search Service it's a matter of less than 30 seconds. Why such a big difference?

For quick keyword-based search, Azure Search Service remains the preferred choice. However, if you require advanced capabilities like semantic search, Azure AI Studio provides powerful tools, but with a longer setup time. Always remember that Azure AI Studio is designed for advanced use cases, and its capabilities might require additional setup time compared to the more straightforward Azure Search Service. Choose the method that best suits your needs!

Since Azure AI Studio is still in preview, there may be optimizations and improvements that have not yet been implemented. As the service matures and moves out of preview, you may see performance improvements.

Azure AI Studio

• It is designed for more advanced use cases, such as RAG. It allows you to create vector indexes for tasks like semantic search and natural language generation.
• It leverages sophisticated algorithms and pre-trained models (such as embedding models) to handle vector-based search. These models require additional processing time during index creation.
• In Azure AI Studio, you can choose an embedding model (e.g., Azure OpenAI) for vector search. These models generate dense vector representations for text, enabling semantic similarity comparisons. The embedding model deployment and initialization contribute to the longer index creation time.
• When creating a vector index in Azure AI Studio, ensure that you have the appropriate permissions and consider the following as a best practice to avoid errors: Avoid using Owner, Contributor, or Azure AI Developer roles at the project level. Assign additional permissions using the recommended methods

Azure Search Service

• It is primarily focused on traditional keyword-based search. It creates simpler indexes optimized for quick retrieval of documents based on keywords.
• It relies on standard inverted index structures, which are faster to build and query.
• Azure Search Service manages physical structures efficiently, allowing for faster index creation.

(It's also not clear how to delete an index from Azure AI Studio Preview)

Using azure portal:

1. Go to the Azure Portal.

Navigate to the search service resource that you want to manage.

Within the search service, select Indexes.

Find the index you want to delete and remove it from there.

Using Delete Index operation via the API: (**test before implementing)

DELETE https://[service name].search.windows.net/indexes/[index name]?api-version=[api-version]
Content-Type: application/json
api-key: [admin key]

Please 'Upvote'(Thumbs-up) and 'Accept' as an answer if the reply was helpful. This will benefit other community members who face the same issue.