Powering AI-Ready MySQL: When MyVector Meets ProxySQL (René Cannaò, Alkin Tezuysal) скачать в хорошем качестве

Powering AI-Ready MySQL: When MyVector Meets ProxySQL (René Cannaò, Alkin Tezuysal)

As modern application architectures increasingly incorporate vector-based retrieval and AI queries, traditional MySQL deployments face challenges when integrating without introducing fragmentation. This session presents a unified architecture that combines MyVector, a native vector extension for MySQL, with ProxySQL as an orchestration layer. This setup creates a single control plane for managing both SQL and vector workloads, enhancing ProxySQL’s routing capabilities for content-aware query classification and dynamic policy enforcement. We’ll discuss how ProxySQL manages vector-aware traffic, isolates resources between transactional and AI workloads, and oversees embedding pipelines while maintaining MySQL’s stability. Attendees will learn architectural patterns and design considerations for safely evolving MySQL into an AI-capable platform, solidifying its role in modern data workloads.