2026 DB&T Oracle Database March Meet-Up
We look forward to having you join us for our upcoming Oracle Database SIG meetup.
These conversations are created with you in mind—your challenges, your insights, and your successes. If there’s a topic you’d like to see highlighted at an upcoming meetup, or if you’ve implemented something innovative or found a creative solution you’d like to share, we’d love to hear from you!
Just reach out to Michaela (michaela.sturgill@questoraclecommunity.org) —we’re always looking to feature voices from the community.
From REST APIs to AI Agents: Managing Oracle GoldenGate Through the Model Context Protocol
Speaker: Bobby L. Curtis, MBA — CTO, Chief AI Architect & Oracle ACE Director, RheoData
Date: March 18, 2026
Abstract
Oracle GoldenGate has long served as the backbone of enterprise real-time data replication, but managing its Extract and Replicat processes, monitoring lag, validating credentials, and troubleshooting failures still demands deep operational expertise and manual interaction with the Administration Service. The idea of using AI to change that equation is not new for this speaker — in May 2024, while preparing for the AIOUG Yatra conference, Bobby Curtis began building the foundations for AI-driven GoldenGate diagnostics by leveraging Oracle Database 23ai Vector Search and Indexes to enable AI to analyze and diagnose GoldenGate environments. That early work proved the concept; the Model Context Protocol (MCP) now provides the standardized framework to make it operational at scale. This session demonstrates how GoldenGate’s RESTful APIs and MCP — an open standard that enables AI agents like Claude and Microsoft Copilot to securely connect to external tools — come together in a working implementation, GoldenGateMCP, that exposes over 200 GoldenGate API endpoints as AI-callable tools across dual transport modes (stdio for local agents, stream able HTTP for platforms like Copilot Studio), with built-in observability through metrics collection, process state tracking, API response caching, and execution tracing that gives operators full visibility into what the AI is doing and why.
More importantly, we will explore what this means for the future of database operations. When a DBA can ask “show me all abended processes in the target environment” and receive an accurate, real-time answer through a conversational interface, we fundamentally change how teams interact with their replication infrastructure. This is not about replacing expertise — it is about making that expertise more accessible, more responsive, and more scalable. We will also look ahead at the roadmap from AI agents toward truly agentic AI, where the system moves beyond responding to questions and begins proactively monitoring process health, detecting failure patterns, and recommending remediation paths autonomously. The foundational pieces — state tracking, operational pattern detection, and system health synthesis — are already in place. The next step is closing the loop from observation to action.
Three Key Takeaways
- MCP Is the Bridge Between AI Agents and Your Infrastructure. The Model Context Protocol provides a standardized, secure way for AI agents to call external tools. Attendees will leave understanding how MCP works at the protocol level — tool registration, transport modes, and authentication — and why it matters for anyone managing complex middleware like GoldenGate.
- GoldenGate’s REST APIs Enable a New Class of Operational Intelligence. Oracle’s decision to expose GoldenGate administration through RESTful APIs created an integration surface that goes far beyond scripts and dashboards. By wrapping these endpoints in an MCP server with built-in metrics, state tracking, caching, and tracing, teams gain an AI-powered operations layer that delivers real-time process monitoring, failure detection, and performance analysis through natural language.
- Conversational Infrastructure Management Is Here — And Agentic Operations Are Next. Managing replication processes through an AI agent is not a future concept; it is a working implementation. Attendees will see how multi-environment support, dual transport modes (local and remote), and portable distribution through .mcpb bundles make it practical to deploy AI-driven GoldenGate management today — and how built-in state tracking, pattern detection, and system synthesis lay the groundwork for fully agentic operations that proactively identify and resolve issues across enterprise environments.
Thank you for joining and for being a valued part of this community. We’re glad you’re here!
