LinkedIn·Tuesday, 18 August 2026·8d ago
Welcome to our series on building the perfect Event-Driven Architecture! We're breaking down the four layers of a production-ready EDA…
Cymo
1,355 followers
Welcome to our series on building the perfect Event-Driven Architecture! We're breaking down the four layers of a production-ready EDA setup and our favorite technologies for each.
First up: the core the transport layer that forms the foundation of your setup. These are the streaming engines and managed cloud services that actually move your data. Much like a car, picking the right engine isn't easy: it depends on your setup, cloud strategy and performance needs. Don't just pick what's popula, the operational overhead can quietly eat your team's time.
Here are the five technologies we rely on for the core, and how we use them:
🔹 Confluent enterprise-grade Kafka without the operational burden. We enforce strict data contracts between teams via Schema Registry and run real-time SQL with ksqlDB.
🔹 RedPanda a lightweight C++ alternative to Kafka built for performance and low latency. We run high-performance clusters with zero external dependencies, even on resource-limited IoT edge gateways.
🔹 StreamNative fully managed Apache Pulsar that handles both queuing and streaming. We use it to offload petabytes of event history to cheap S3 storage while keeping it fully queryable.
🔹 Amazon Web Services (AWS) MSK managed open-source Kafka, native for teams already on AWS. We migrate self-hosted clusters to the cloud with minimal downtime and manage access with standard IAM roles.
🔹 Aiven a managed platform for multi-cloud strategies. We spin up a cloud-agnostic stack in one console: Kafka for transport, Flink for processing, ClickHouse for analytics.
Not sure which engine fits your infrastructure? Let's figure it out together → http://cymo.eu
#EventDrivenArchitecture #ApacheKafka #DataStreaming #CloudInfrastructure #Confluent
♥ 13↻ 2
View on LinkedIn Cross-referenced
Related on the wire
Stop reverse-engineering your business. It's time to model reality. For decades, we've hidden business facts in database tables. But what…
“Kafka is durable by design, so I don't need backups.” WRONG! ❌ A lot of people forget that the biggest threat to their stability isn't…
"Kafka is durable, so I don't need backups." This is a common—and dangerous—assumption. In his latest article, our expert Pieter explains…
"...data teams end up spending 80% of their time cleaning and preparing messy exports from operational systems. It's wasted energy.…
Who is responsible for building state in your Event-Driven Architecture? If the answer is "every consumer," you're creating a system that…
If you’re just starting out with Event-Driven Architecture, you might think you just have to pick a streaming broker and you’re done. Well,…