Why is Clean Architecture so Popular?

You’ve probably noticed many videos and blogs that somewhat explain what Clean Architecture is and show you how to use it. So its Clean Architecture is popular, but should it be? Should you be using it? Here’s why I think it’s popular, the problems it addresses, and some aspects that almost nobody ever mentions.

YouTube

Check out my YouTube channel, where I post all kinds of content accompanying my posts, including this video showing everything in this post.

Clean Architecture

Clean Architecture
https://blog.cleancoder.com/uncle-bob/2012/08/13/the-clean-architecture.html

As a quick primer, what is clean architecture? Well, it’s a way to manage coupling. Specifically, in this diagram, you can see how the outer parts of the circle reference the inner parts of the circle. The dependencies between layers are pointing in a single direction inward.

It’s about managing coupling.

As an example, with the Clean Architecture template for .NET/C#, the project structure and dependencies are as follows.

Clean Architecture Direction of Dependencies

The top (outer layer), called WebUI, is ASP.NET Core. It references an Infrastructure project that contains the entity framework DBContext and other concerns. The WebUI and Infrastructure reference the Application project, which contains the interfaces for implementations in the infrastructure and any application-level code, such as commands, queries, and handlers. Finally, the application project references the Domain project, which contains (or should) your domain models and business logic.

Sounds great. Separation of technical concerns. But why?

Coupling

degree of interdependence between software modules

ISO/IEC/IEEE 24765:2010
Systems and software engineering — Vocabulary

Big ball of Mud

There are two forms of coupling Clean Architecture addresses. Afferent and Efferent.

Efferent Coupling: Who do you depend on? From the perspective of the Domain project, who does it depend on? Nothing.

Afferent Coupling: Who depends on you? From the same perspective of the domain project, which projects depends on it? The Application Project.

This is about stability. Because the Domain project has no dependencies, nothing can force it to change. All our business logic is isolated and cannot be forced to change because of a change within the infrastructure project or any other project. The reverse is true for WebUI. Changes we make in the infrastructure or Application could force us to make changes in the WebUI.

Do you need Clean Architecture?

It would be best if you asked yourself a few questions. What is the size of the application? Do I have complex domain logic? Do I need to control coupling?

Clean architecture is about forcing a direction of dependencies. In .NET, projects were used in the template above to force the separation. However, you do not need separate projects. Coupling is the dependence between types. If you merged the template into one project, you still have the same degree of coupling.

Prescription

Do not use Clean Architecture as a prescription or template. Understand that you’re trying to manage coupling. It doesn’t need to be by projects. However, it can be to help with physical separation. It doesn’t need to be those exact layers. It’s not a prescription.

Large System

You should consider decomposing it into logical boundaries if you have a large system. What’s a large system? Something that takes a team of developers, possibly years to develop. I’ve covered this in many different blog posts and videos. Check out my post Microservices gets it WRONG defining Service Boundaries and Should you use Domain Driven Design? where I talk about logical boundaries. Logical boundaries are about grouping a cohesive set of capabilities within your system. It allows you to decompose a large system into smaller subsystems.

Logical Boundaries

Why does this matter? When you break up a large system into smaller parts, you’ll realize that not all parts provide the same value. While all the boundaries are important, some are more in a supporting role and often built around CRUD (Create-Read-Update-Delete). This is also very similar if you’re creating a smaller app that may take a couple of weeks or months to develop.

If you have no domain logic, do you need to all the same layers as another part of your system that is at the core of the solution space and contains complex business logic? No.

Clean Architecture within logical boundaries

This is why it’s not a prescription or template. Each boundary within a system has different concerns. If you don’t have any business rules, you have an underlying data model. Or perhaps you only have a dozen or so routes/endpoints that have data access. Do you need to add an abstraction to data access in that case? What if your database changes? Then change the 12 or so routes/endpoints!

Clean Architecture

Clean architecture is about coupling. There’s no prescription for the layers you define or how you define the coupling. You don’t need to define layers by projects. It’s about the direction of dependencies between types. Afferent and Efferent coupling are what define the stability of each layer. Do you need stability in a particular layer? Then maybe consider isolating it.

Join!

Developer-level members of my YouTube channel or Patreon get access to a private Discord server to chat with other developers about Software Architecture and Design and access to source code for any working demo application I post on my blog or YouTube. Check out my Patreon or YouTube Membership for more info.

Follow @CodeOpinion on Twitter

Software Architecture & Design

Get all my latest YouTube Vidoes and Blog Posts on Software Architecture & Design

That’s NOT an Aggregate in Domain Driven Design

Are you frustrated that you have to open multiple files across multiple layers to make what seems like a simple change? One of the culprits for this is following structure and templates that apply patterns or concepts to solve problems you might not have. One typical case of this is using aggregate from domain drive design. In this video, I’ll give examples of where an aggregate can make sense and where it’s not and adds useless indirection.

YouTube

Check out my YouTube channel, where I post all kinds of content accompanying my posts, including this video showing everything in this post.

Useless Indirection

The idea for this video/blog came from a common I received on a YouTube video on my channel where I was talking about indirection.

The sentiment of this comment is all too common. I have many similar comments and have conversations with developers all the time about this. One of the culprits to this is applying patterns or concepts that are solutions to problems you don’t have in a given context.

One such pattern is the usage of an Aggregate from Domain Driven Design. The purpose of an aggregate is to create a consistency boundary. Unfortunately, the way it’s often explained more illustrates it as an object model or hierarchy.

Aggregate Domain Driven Design

The stereotypical example is to model a shopping basket. You would have a basket that would have many basket items. Many think this is an aggregate because you cannot have a basket item without a basket. In this case, this would be the aggregate, and the Basket would be the aggregate root.

Typically you’d then use a Repository to save and fetch the aggregate out, only exposing the aggregate root (Basket) to consumers.

Aggregate Domain Driven Design

But does this need to be an aggregate?

Most commonly, aggregates are often incorrectly used to model an object/data hierarchy and to old domain logic, which I often think is a more trivial validation than complex domain logic.

However, an aggregate is about creating a consistency boundary. It’s not about modeling a hierarchy.

Do you need consistency within this aggregate?

Useless Setters

Here’s a made-up example of an aggregate based on a sample I found on GitHub.

This is a simplified example. However, you can see two methods for setting the Name and the Price of this Entity. There is also some logic for setting the price: the price must be greater than zero. To do this, it’s using a specification.

What value does the specification serve? What value do the SetName and SetPrice have? None.

The SetName method is just setting the underlying Name property. It’s useless indirection.

The SetPrice contains some validation logic, which is nice. However, the separate ProductNegativePriceSpecification is useless indirection. The SetPrice is also putting our entity in an invalid state even though it’s throwing. The caller could catch the exception and carry on.

We could just put the conditional check directly in the SetPrice method. But we can also use value objects and types to enforce a valid value directly from the caller.

Now, what value do the SetName and SetPrice have? Zero value. They are just setting the underlying properties. We’ve enforced our product price when the caller needs to construct a ProductPrice type.

We don’t have an aggregate (root). We have a data model with useless setters. Remove the SetPrice and SetName, then set the properties directly from the calling code.

Consistency Boundary

So when do you need an aggregate? Well, here’s an example of an Order Aggregate (root)

This slimmed-down version of the Order Aggregate Root illustrates what’s important. When we add an order item, we do it through the aggregate root (Order) because we want to only have a single unique product per order. Also, if we have a discount for the product, we want to use the discount with the greatest value. This is a consistency boundary. We need an aggregate and all operations to go through the root to perform this logic. We don’t want random data access code or transaction scripts managing order items. This gives us consistency.

Lastly, in the SetStockconfirmedStatus method, we’re making a state change, but we’re also publishing a domain event OrderStatusChangedToStockConfirmed. Other parts of our system likely rely on this event when that state changes. We must always publish this event when the order status changes to StockConfirmed. Again, consistency on state change and publishing an event.

Aggregate or Data Model

If you need a consistency boundary, use an aggregate and aggregate root. You’re not getting any of the benefits if you have a data model with just setters. Don’t add useless indirection. Just use a data model with transaction scripts.

Join!

Developer-level members of my YouTube channel or Patreon get access to a private Discord server to chat with other developers about Software Architecture and Design and access to source code for any working demo application I post on my blog or YouTube. Check out my Patreon or YouTube Membership for more info.

Follow @CodeOpinion on Twitter

Software Architecture & Design

Get all my latest YouTube Vidoes and Blog Posts on Software Architecture & Design

McDonald’s Journey to Event-Driven Architecture

McDonald’s uses Event-Driven Architecture! Luckily for us, they’ve written a couple of blog posts providing some details of their journey into event-driven architecture. I’m going to go a bit deeper by providing my thoughts on how their system works and why they are doing it so that it can give you some ideas about your systems.

YouTube

Check out my YouTube channel, where I post all kinds of content accompanying my posts, including this video showing everything in this post.

McDonald’s Event-Driven Architecture

It’s always interesting to see companies post details of the architecture of various systems they have. It can be insightful to see what they are doing, why, and their challenges. McDonald’s posted behind-the-scenes and how-it-works blog posts detailing their journey to event-driven architecture. More specifically, it’s not that they are new to event-driven architecture but rather have a standardized way to implement it with distributed teams of developers with different skill levels.

McDonald's Event-Driven Architecture

There are many different components to their platform. Their infrastructure is within AWS, and they use MSK (Managed Streaming for Kafka) along with ECS, DynamocDB, and API Gateway.

Here’s how everything works together.

Schema Registry

One of their challenges was related to data quality. Likely because there was no set definition (schema) for data within events. If multiple producers produce the same event type, they might not be composing them exactly the same. I believe an event should have a single publisher, the owner of that schema, to avoid this issue. However, this could be applicable in a message-driven architecture that’s also using queues and commands.

Producers at startup use a custom SDK that retrieves all the event schemas from the registry. This allows the producer to validate the event being produced against the schema.

If validation passes, the producer can publish this event to the appropriate Kafka topic using the SDK at this point.

As you can expect, on the consumer side, the same thing occurs. Consumers at startup use a custom SDK that retrieves all the schemas from the registry, just like the producers do.

Then the consumers can process messages from the Kafka topics and understand how to deserialize them from the schema and version of the schema.

Everything within any Kafka topic should be valid based on all the schemas (versioned) within the registry. Data quality issues are solved!

Validation

Of course, not everything goes through the happy path. What if a producer tries to publish an event, but it fails to validate against the schema? The producer then publishes the message to a Dead Letter Queue. Kafka isn’t a queue, so this is a Dead Letter Topic.

Producer to DLQ

Once a message is in the “DLQ” there needs to be a way to view, modify and fix the event so it can be re-published to the correct topic.

For this, an Admin/Utility UI provides this functionality for them.

Reliable Publishing

The second failure that can occur is failing to publish to Kafka (MSK). Anyone getting involved in Event-Driven Architecture is bound to run into this. It would be best if you had consistency between making state changes to your business data and publishing your event. When events become critical to your system and possibly workflows, you need guarantees that you publish the relevant events when you make some state change to business data.

Mcdonald’s chose to use DynamoDB to persist any events that cannot be published to Kafka. This means their Publisher SDK will fallback to storing the event data within DynamoDB if it cannot publish to Kafka.

Using a fallback to some durable storage is a common approach. However, the Outbox Pattern is another common solution. I discussed this and other common issues in a post about the 5 Pitfalls of EDA.

Once the event data is in DynamoDB, they use Lambda to pull it from DynamoDB and then retry and publish it to Kafka. I’d assume they have different retry intervals/backoffs.

Lambda Retry

Gateway

Lastly, if you’re integrating with 3rd parties or even within a large organization, you’ll need to have them publish events. However, they won’t have direct access to your SDK and Kafka. For this, they use API Gateway as an HTTP interface to convert HTTP requests that will communicate with the Producer that has the SDK and can publish to Kafka.

Event Gateway

That way, we go through the same validation against the schema in the registry just as if any of our client code is using the producer SDK. This allows external 3rd parties to publish events without using our SDK directly. We can instead have them use our Event Gateway (HTTP API).

Technical Blog Posts

I love when companies have technical blog posts that give insights into their architecture and design. It’s hard to know the full context, but seeing how they solve these issues they run into is interesting. Companies face many common issues when using Event-Driven Architecture, but all have unique constraints.

If you have any recommendations for other technical blog post analyses, please let a comment!

Join!

Developer-level members of my YouTube channel or Patreon get access to a private Discord server to chat with other developers about Software Architecture and Design and access to source code for any working demo application I post on my blog or YouTube. Check out my Patreon or YouTube Membership for more info.

Follow @CodeOpinion on Twitter

Software Architecture & Design

Get all my latest YouTube Vidoes and Blog Posts on Software Architecture & Design