Build a semantic layer for agentic AI on AWS with Denodo and Amazon Bedrock AgentCore
aws ·
This article details the construction of a semantic layer for agentic AI on AWS using Denodo Platform and Amazon Bedrock AgentCore. It extends a previous customer 360 pattern by incorporating an on-premises PostgreSQL database alongside Amazon Aurora and Amazon Redshift. The solution uses Denodo’s Virtual DataPort to federate these disparate sources into a unified view without moving data. An agent hosted on AgentCore queries this layer via the Model Context Protocol (MCP), with identity propagation ensuring secure, role-based access control. Global Security Policies enforce row and column filtering at query time, protecting sensitive information like PII. The architecture supports two deployment options: the fully managed Denodo Agora or self-managed installations on EC2 and EKS. This approach allows foundation models to reason over business-defined metrics rather than raw schema, improving accuracy and governance.
aws