Kioxia Integrates AiSAQ Vector Search Technology Into Milvus

KUALA LUMPUR, Dec 17 (Bernama) -- Kioxia Corporation announced that its KIOXIA AiSAQ, an Approximate Nearest Neighbor Search (ANNS) software technology, has been integrated into the open-source vector database Milvus beginning from version 2.6.4.

With the integration, Milvus users can leverage KIOXIA AiSAQ solid-state drive (SSD)-optimised vector search capabilities, providing developers and enterprises with a practical and cost-efficient way to scale artificial intelligence (AI) applications, according to the company in a statement.

Kioxia noted that the AI industry is shifting from building large foundation models to deploying scalable and cost-effective inference solutions that address real-world use cases.

Retrieval Augmented Generation (RAG) plays a central role in this transition, and KIOXIA AiSAQ was developed to help the open-source community adopt SSD-based vector architectures to overcome limitations associated with scaling dynamic random-access memory (DRAM).

First announced earlier this year, KIOXIA AiSAQ is an open-source software technology designed to significantly improve vector scalability by storing RAG-related database elements on SSDs while maintaining high-quality vector search accuracy.

The integration into Milvus is expected to enhance adoption within the open-source ecosystem and support developers in building faster and more efficient AI applications.

Kioxia said the collaboration enables a new class of scalable, cost-efficient vector search solutions to meet the growing demands of modern AI workloads.

A global memory solutions provider, Kioxia focuses on the development, production and sale of flash memory and SSD products.

-- BERNAMA