Enhancing Oracle HeatWave with in-database LLMs for reduced infrastructure costs

Spread the love



Oracle is updating its Heatwave cloud data analytics service, previously known as MySQL Heatwave, with new AI-focused generative features. The rebranding to HeatWave signifies a broader offering beyond just MySQL support, including HeatWave Gen AI, HeatWave Lake House, and Heatwave AutoML. These updates were showcased at Oracle’s CloudWorld conference in September 2023, introducing features such as a large language model (LLM) interface, Vector Support for Store, Heatwave Chat, and AutoML for HeatWave Lakehouse.

The newly introduced HeatWave Gen AI combines these enhancements and is now available at no extra cost. Analysts note that Oracle is the first database vendor to incorporate LLM support within a database, aiming to reduce infrastructure costs. By utilizing smaller LLMs like Mistral-7B and Goal Flame 3-8B, businesses can save on memory consumption and use CPUs instead of GPUs, making it a cost-effective solution.

The integration of LLMs in databases enables various applications such as data search, content generation, and Recovery Augmented Generation (RAG) with HeatWave’s Vector Store. Additionally, HeatWave Gen AI features Oracle’s Vector Store for scalable vector processing, allowing faster query processing without the need to move data to a separate vector warehouse.

HeatWave Chat, a new addition to HeatWave Gen AI, offers a visual code companion for mysql Shell, enabling developers to interact with HeatWave GenAI through natural language or SQL queries. This feature includes Lakehouse Navigator, facilitating the selection of files from object storage to create a new vector store, enhancing the user experience for developers and analysts.

Oracle’s focus on AI-driven capabilities within HeatWave Gen AI aligns with current industry trends, as other database providers have started incorporating vector capabilities into their products. By supporting ventures such as generative AI and AI-based use cases directly within the database, Oracle aims to streamline processes and improve query performance.

Overall, the introduction of HeatWave Gen AI signifies a significant step forward for Oracle in enhancing its Heatwave service with advanced AI features, such as LLM integration, Vector Store, and scalable vector processing. These updates are poised to benefit businesses by reducing costs, improving query efficiency, and providing a more user-friendly interface for developers and analysts.

Article Source
https://www.infoworld.com/article/3715583/oracle-heatwave-s-in-database-llms-to-help-reduce-infra-costs.amp.html