PlanetScale, a cloud-based database service developed from MySQL, has announced the separation of the MySQL project in order to enhance the Vector Database feature for storing data from artificial intelligence. This feature was previously announced by Oracle as part of the MySQL HeatWave, a cloud service. However, this move suggests that Oracle will focus on storing this feature for cloud services.
Storing vectors in any database is feasible; however, the key aspect is the indexing process to ensure quick data retrieval. PlanetScale has indicated that their new project will utilize the Hierarchical Navigable Small World (HNSW) algorithm.
The vector feature has become a significant aspect of database software this year. Database software has been incorporating this feature, whether it is directly implemented within the database or developed as an extension to support it. Examples of databases that support this feature through projects like pgvector include Oracle, MongoDB, and PostgreSQL.
TLDR: PlanetScale, a MySQL-based cloud database service, is separating the MySQL project to enhance the Vector Database feature, which stores data from AI. They will utilize the HNSW algorithm. This vector feature has gained importance in database software, with various databases incorporating it.