Planning
There are several planning considerations when using vector search. Astra DB makes it easy with the following capabilities:
-
Vector Search requires scale. A typical measurement is ten times the scale at one-third of the cost. Astra DB is a serverless database, making scaling up easy.
-
Astra DB balances the need for faster search by using more connections to offset longer compaction times and less effective caching.
-
Cache settings are applied per
hnsw
index segment. Hierarchical Navigable Small World (HNSW) graphs are among the top-performing indexes for vector similarity search.