The Serverless (Vector) database is designed for vector search applications, such as Generative AI (GenAI), semantic search, and geospatial search. The Serverless (Vector) database stores vector embeddings alongside non-vector tabular data. If you are building Retrieval Augmented Generation (RAG)-based applications, select the Serverless (Vector) database. A few examples of RAG applications include:

The Serverless (Non-Vector) database is designed for generic applications, such as content management, user authentication, and transactional applications.

As technology evolves, so does the terminology used to describe it. The Cassandra® community describes the database components with these terms: cluster, keyspace, table, and column. A keyspace is a container for tables, similar to a schema in a relational database. As language for developers has shifted, this has led to confusion among users and developers. "Namespaces," "collections," and "documents" are more familiar terms for developers. Depending on the database and features you are using in Astra Portal, you might see different terminology used.

This table shows the terminology used in Astra Portal for each database type:

The Data API and clients for Astra DB Serverless use namespace. The Astra Portal uses keyspace in some places, but namespace when interacting with your database through the Data API or clients.

Astra DB Serverless supports multiple regions across Amazon Web Services, Google Cloud, and Microsoft Azure.

The following tables describe the supported regions for each provider, including location, pricing, and supported database types.

DataStax adds new regions based on customer demand. If your preferred region isn’t listed here, you can submit a support ticket or send us a message through live chat in the Astra Portal.