Scalar Inc. has released a new version 3.15 of ScalarDB, a general-purpose HTAP engine that virtually integrates multiple, heterogeneous databases. This version adds an attribute-based access control mechanism that controls access at the fine data unit (record), and also adds a new vector store as a database that can be managed by ScalarDB. These functional enhancements make it easier to build applications that require high security, as well as large-scale language models (LLMs) and retrieval-augmented generation (RAGs). ScalarDB can now be used in a wide range of industries and applications, from security-sensitive industries such as finance to advanced businesses that use AI.
ScalarDB 3.15 adds a “vector store” to the types of databases that ScalarDB can handle, making it easier to add and search vector data through ScalarDB’s vector interface. This enhancement makes it easier to use ScalarDB to realize advanced business using AI. For example, processes such as building Retrieval-Augmented Generation (RAG) in large-scale language models (LLMs) can all be done through ScalarDB.
SOURCE: PRTimes