When building a distributed system, developers are often faced with something known as the dual-write problem. It occurs whenever the system needs to perform individual writes to separate systems that can't be transactionally linked. This situation creates the potential for data loss if the developer isn't careful. However, techniques such as the Transactional Outbox Pattern and Event Sourcing can be used to guard against the potential for data loss while also providing added resilience to the system.

Tabs or spaces? Merge vs. rebase? Flink SQL vs. KStreams? Let’s Settle This is powered by a new Kafka JavaScript client from Confluent: confluent-kafka-javascript (early access). Find out how Lucia used it to make the website in the video above.

The headless data architecture. Is it a fad? Some marketecture? Or something real? In this video, Adam Bellemare takes you through the basics of the headless data architecture and why it’s beginning to emerge as its own respective pattern. Driven by the decoupling of data computation from storage, the headless data architecture provides the basis for a modular data ecosystem. Stream your data for near real-time low latency use cases, or convert it to an Iceberg table for analytical use cases.

In this video, you will see an example of how Tributary bank uses asynchronous events to enrich its domain and protect its fraud detection system from failures.

In this video, we discuss one way a business can approach decomposing a monolith using a series of clearly defined steps and robust monitoring.

Join Lucia Cerchie in a coding walkthrough, bridging the gap between REST APIs and data streaming. Together we’ll transform the OpenSky Network's live API into a data stream using Kafka and Flink SQL. Not only do we change the REST API into a data stream in this walkthrough, but we clean up the data on the way! We use Flink SQL to make it more readable and clean, and in that way we keep more of the business logic away from the client code.

In this video, Wade explores the process of decomposing a monolith into a series of microservices. You'll see how Tributary bank extracts a variety of API methods from an existing monolith. Tributary Bank wants to decompose its monolith into a series of microservices. They are going to start with their Fraud Detection service. However, before they can start, they first have to untangle the existing code. They will need to define a clean API that will allow them to move the functionality to an asynchronous, event-driven microservice.

Learn how the Data Portal and Apache Flink in Confluent Cloud can help developers and data practitioners find the data they need to quickly create new data products.

How do you prevent hallucinations from large language models (LLMs) in GenAI applications? LLMs need real-time, contextualized, and trustworthy data to generate the most reliable outputs. Kai Waehner, Global Field CTO at Confluent, explains how RAG and a data streaming platform with Apache Kafka and Flink make that possible.

How do we know whether Event-Driven Microservices are the right solution? This is the question that Tributary Bank faced when they looked at modernizing their old fraud-detection system. They were faced with many challenges, including scalability, reliability, and security. Some members of their team felt that switching to an event-driven microservice architecture would be the magic bullet that would solve all of their problems. But is there any such thing as a magic bullet? Let's take a look at the types of decisions Tributary Bank had to make as they started down this path.