Unifying Middleware: How EDA Connects the Puzzle Pieces of Integration

At EDA Summit, François Rivard—co-founder of Astrakhan Consulting—delivered a wide-ranging and insightful talk on how event-driven architecture (EDA) is becoming the missing link in middleware integration strategies. With deep client experience across industries and a strong partnership with Solace, Astrakhan has seen firsthand how real-time data movement is transforming integration—from industrial systems to AI agents.

This post summarizes a talk delivered by François Rivard at EDA Summit 2024.

Here’s a 1-minute summary, and you can watch the full presentation either below or at EDASummit.com.

Why EDA Matters

Originally introduced to bridge operational technology (OT) and information technology (IT) in Industry 4.0 contexts, EDA has since evolved into the real-time foundation of modern integration strategies. Rivard described it as a uniquely powerful enabler of agility and unification—allowing disparate systems to interact in a coordinated, real-time manner. By supporting synchronous and asynchronous patterns across domains, EDA makes it possible to integrate business applications, democratize data sharing, and even shift between centralized and decentralized governance models depending on organizational maturity.

EDA is the missing link that unifies every block of integration across complex systems.”

This flexibility is key. With EDA, organizations can adopt a layered approach to data management, balancing autonomy and control while enabling data mesh and data product platforms. And while many associate event brokers solely with microservices, Rivard emphasized that EDA’s utility extends far beyond that. It is increasingly the nervous system of enterprise integration—spanning everything from legacy systems and APIs to modern AI agents that require timely, contextual inputs to function effectively.

Key takeaway:
EDA is not just for microservices — it’s the real-time backbone that unifies everything from legacy systems to AI-powered agents.

How Astrakhan Helps Implement EDA

Rivard outlined an architectural pattern that Astrakhan commonly recommends and helps clients implement using Solace Platform. At the center is a three-tier broker mesh: one broker sits in the front office, one in the back office, and a core broker unifies the two. This structure allows for scalable and resilient real-time messaging across business units, environments, and technologies.

Rather than transforming data inside the broker itself, Astrakhan advocates an adapter-driven approach. Applications send and receive data in canonical formats via adapters, ensuring that the event broker focuses on routing while transformation happens at the edges. This micro-integration model supports a clean separation of concerns and aligns with modern principles of data governance.

You don’t replace iPaaS or APIs with EDA—you complement them to build a truly unified platform.”

Astrakhan also helps clients integrate EDA with other styles of middleware. For example, iPaaS platforms handle orchestration and complex data transformations, while API gateways manage synchronous connections and external exposure. These components aren’t competitive—they’re complementary. Unified through the Solace Platform, they form the backbone of what Rivard called a “hybrid integration platform.”

Key takeaway:
Complement, don’t consolidate — each middleware tool brings unique strengths that shine when unified through event-driven architecture.

Real-World Impact

Astrakhan’s client work illustrates the value of this unified approach. A luxury retail group, for example, used real-time messaging to integrate its manufacturing systems with business operations and online boutiques, enabling consistent end-to-end customer experiences. Other clients have adopted data mesh architectures using EDA to allow departments like finance, sales, and customer service to maintain ownership of their data while publishing it in shared, canonical formats.

From manufacturing to boutiques to AI agents, EDA connects every part of the business in real time.”

Perhaps most forward-looking is the use of EDA to connect AI agents. By delivering real-time events to LLMs, machine learning models, and autonomous agents, Astrakhan has enabled clients to build what Rivard calls a “context mesh.” This architecture allows AI systems to behave intelligently based on the current state of the business—bringing real-time responsiveness to AI-powered workflows.

Key takeaway:
EDA makes AI smarter — feeding context to agents in real time elevates your AI from reactive to proactive.

Final Thoughts

Rivard closed his talk by emphasizing that tools alone aren’t enough—success with EDA depends on organizational maturity and governance. He described how a unified governance portal can serve as the nerve center of integration, bringing together APIs, events, and data products in a single interface. Such a portal doesn’t replace native capabilities like the Solace Event Portal or API product catalogs, but rather federates them into a coherent platform for design-time planning and runtime observability.

Treat integration like a product—designed with common data formats, governed centrally, and reused across domains.”

He also underscored the importance of consistent topic taxonomies and common data formats across integration initiatives. These architectural decisions ensure compatibility and reuse, accelerating time to market and reducing duplication. When integration accounts for up to half the effort in major IT projects, this kind of reuse becomes a competitive advantage. Ultimately, Rivard encouraged organizations to treat integration like a product—one that evolves through reuse, federation, and cross-domain collaboration.