# 21.05 Event-Driven Services

> Plan, build, and deploy event-driven architectures, with Kafka as the working example.

Format: 3 days · instructor-led
Builds on: 21.03

Most service architectures start synchronous: one service calls another and waits. That's simple until a slow dependency stalls everything upstream, or a new service needs data locked inside another team's database.

Event-driven architecture turns that around. Services announce what happened, and anyone who cares can react, now or later. This course covers when that's worth the trouble and when it isn't, how to design events and contracts that survive versioning, how to produce, consume, and process event streams, and how to bring events into systems that weren't built for them. Apache Kafka is the working platform, with just enough of its architecture to make sense of the rest; the patterns apply to most streaming platforms.

## Who it's for

.NET developers familiar with service architectures, ideally after [Web API 200](/21.03-web-api-200/). Some understanding of RPC (HTTP, gRPC) and pub/sub messaging helps. Some familiarity with Java and SQL helps too, but isn't required.

## How it runs

Three days, live and online: lecture, theory, design labs, coding exercises, and group discussion. It's synchronous, and you'll be asked to participate fully during class time.

This course is not intended to be a thorough introduction to any particular streaming platform, but will utilize Apache Kafka, with aspects from the Confluent Kafka platform. While securing event streams will be discussed, no particular guidance will be given in this area as it varies so much by application and environment (regulatory, etc.).

## What you'll be able to do

- Recognize the problems event-driven architecture solves, and the antipatterns it invites.
- Design events and their contracts with explicit schemas, and version them without breaking consumers.
- Build services that produce and consume events, and know which delivery guarantees you're getting.
- Write stateless and stateful stream processors.
- Use event streams and change data capture to free data from a central database, and decide which services own it.
- Introduce events into existing systems, including legacy ones.
