Event-Driven Architecture on AWS: Everything Reacts, Nothing Waits
I recently spoke at BlablaConf 6.0 Day 4 (Cloud, Containers & Infrastructure) about Event-Driven Architecture on AWS. In this post, I recap the talk: core concepts, AWS building blocks, and a real-world Taco Bell case study.
A few months ago, I had the opportunity to speak at BlablaConf 6.0 β Day 4: Cloud, Containers & Infrastructure π²π¦, the annual conference organized by the GeeksBlabla community, where talks are given in Moroccan Darija.
I gave a talk about Event-Driven Architecture (EDA) on AWS, under the motto: βEverything reacts, nothing waits.β
π₯ Want to watch the full episode? Itβs available on YouTube (in Darija): BlablaConf 6.0 Day 4 - Cloud, Containers & Infrastructure
Hereβs a written recap of what I covered.
What is Event-Driven Architecture?
Event-Driven Architecture is an architectural pattern based on asynchronous communication between services through events, instead of direct API calls.
With this approach, our services remain:
- Independent and loosely coupled
- Extensible β new services can join without touching existing code
- Scalable β each component scales on its own
In short: everything reacts to events, and nothing waits. There are no services blocking each other waiting for responses, like weβre used to with REST APIs. This is exactly why EDA is such a great fit for cloud-based systems.
The Building Blocks of EDA
The whole architecture revolves around events β a signal that something happened in the application, like a user interaction or a state change (for example, the creation of a new user).
The architecture is made of three main components:
- Event Producer (Publisher) β any interface or service whose role is to create events and send them to the event broker.
- Event Broker β the central piece that connects producers and consumers. It receives incoming events, filters them, and routes them to the right consumers.
- Event Consumers β services that listen for events from the broker and run their own logic (for example, a notification service sending an email with the userβs credentials).
ββββββββββββ events βββββββββββββββββ events ββββββββββββββ
β Producer β ββββββββββββββΊ β Event Broker β ββββββββββββββΊ β Consumer β
ββββββββββββ βββββββββββββββββ ββββββββββββββ
β β β β
β βΌ βΌ β
β (filter & route) β
β β
ββββ more producers ββββββββββββββββββββββββ more consumers βββ
Why is this so powerful?
- If a producer goes down, consumers keep working normally β they simply consume the events still sitting in the broker.
- If a consumer goes down, producers keep working normally and keep sending events to the broker.
This also makes integrating new services trivial: any new service just follows the events it needs from the event broker, without us having to modify the existing code. The result is an architecture that is extensible and well-decoupled compared to classic API calls β which is why EDA fits cloud environments like AWS so well, since many AWS services are designed to natively support event-driven communication.
EDA on AWS
Letβs map these concepts to actual AWS services:
| EDA Role | AWS Service | Notes |
|---|---|---|
| Producers | S3, DynamoDB, API Gateway, Lambda | Emit events when something happens: file upload, database update, API request⦠|
| Event Broker | Amazon EventBridge | Receives events and routes them based on filters, rules, and event schemas |
| Event Broker (queueing) | Amazon SQS | When events need to be queued and processed by many services |
| Event Broker (broadcast) | Amazon SNS | When an event must be broadcast to many consumers at once (email, SMS, analyticsβ¦) |
| Consumers | AWS Lambda, Step Functions | React to events with business logic and scale automatically with the number of events |
Use Cases
EDA shines in many scenarios:
- IT automation
- Microservices communication
- External SaaS integration
To understand this better, letβs look at a real-world example: Taco Bell, which needed to integrate food delivery applications with its point-of-sale systems using event-driven architecture.
The Taco Bell Flow
- A customer places a Taco Bell order on a delivery app like Glovo (an external SaaS).
- Glovo validates the order and sends it to Taco Bellβs APIs.
- API Gateway receives the order API request and transforms it into an event.
- EventBridge picks up these events and routes them to the business workflow.
- AWS Lambda functions process the orders and apply the business logic.
- Finally, the orders arrive, ready, at Taco Bellβs point-of-sale systems.
Glovo (SaaS) βββΊ Taco Bell APIs βββΊ API Gateway βββΊ EventBridge βββΊ Lambdas βββΊ POS
order validation event routing process system
Wrapping Up
Event-Driven Architecture brings decoupling, resilience, and scalability to distributed systems β and AWS gives you all the building blocks (EventBridge, SQS, SNS, Lambda, Step Functions) to build it natively.
If you want to dive deeper, check out the resources below:
- π₯ BlablaConf 6.0 Day 4 - Cloud, Containers & Infrastructure (my talk, in Darija)
- π BlaBlaConf β the annual conference by the GeeksBlabla community
- π AWS - Event-Driven Architecture
Thanks to the GeeksBlabla team for organizing BlablaConf, and to everyone who watched the episode. Thalaaaaaaw! π