We now have Starter Projects for Django, Flask, Nest, and Nuxt! What should we build next?

All Posts

Microservice orchestration or choreography: Which one do you need?

Lee Brandt - 2022-06-30

The adoption of microservice architecture came into fruition due to advancements in technology (such as GPU) and the exponential rise in data generation. At present, microservices are a well-established architecture framework that allows organizations to deliver large-scale applications rapidly. This blog post will cover the difference between microservice orchestration and microservice choreography, enabling microservices to communicate and coordinate. 

Here’s what you’ll learn: 

Owing to their packet size, microservices are easily implementable and reusable. They ensure rapid delivery, save resources, and are highly scalable. There’s no reason not to adapt to a microservice architecture. However, these small moving components (microservices) must be able to communicate and work together to develop applications and software. Microservice orchestration and choreography enable effective implementation and interaction of numerous microservices. 

Microservice orchestration

An orchestra’s most prominent feature is how each musician plays an instrument while also ensuring that they stay in sync with the other musicians. The conductor of the orchestra is responsible for administering the so-called sync. 

Like a musical orchestra, microservice orchestration enables the execution of a query or process by allowing each microservice to implement tasks assigned by the orchestrator (conductor). While each microservice implements the task and reports the progress, the orchestrator stores the required information within its storage space (common storage). This accounts for the “sync” component of the orchestra. 

By following the commands of the orchestrator, any software transaction or website query is executed efficiently. The role of each microservice is to execute the task or workflow. And the role of the orchestrator is to track the microservices: their status, progress, and so on. This makes orchestration a centralized workflow execution mechanism with the orchestrator playing the “godlike” entity. 

For example, consider a website transaction that entails ordering many items. In such a situation, the orchestrator receives the task and queues it to assign it to microservices. Each microservice works on responding to some API request by querying data. The orchestrator collects the data and sends it to the client. Thus, the orchestrator manages the overall communication and interaction of the workflow. 

Microservice choreography

Like most things in nature, there exists a dichotomous alternative to the workflow that enables communication (interaction) between microservices. As opposed to the orchestration workflow, microservice choreography follows a decentralized mechanism. 

In choreography, the orchestrator does not exist. Instead, the microservices follow an asynchronous approach that entails executing queries when an event is raised. 

The asynchronous approach allows microservices to cooperate and fulfill an event. In this approach, an event queue (message queue) acts as a placeholder for client requests. Only the relevant microservices receive the message and then implement the task listed in the message. Once the task is done, the microservices respond with a success or failure token. This way of communication enables the microservices to coordinate without requiring an orchestrator. 

Choreography enables microservices to work autonomously, removing any issues caused by dependencies. This is important because it allows for the parallelization of tasks or events. 

Orchestration vs. choreography: What’s the difference?

Both microservices orchestration and microservices choreography have their pros and cons. Let’s now cover the differences between these communication workflows and patterns. 

Overall approach

In microservices orchestration, the orchestrator is the brain, or logic component, that assigns tasks to the microservices. On obtaining a transaction request or a query, the orchestrator decides on the workflow and distributes tasks to the microservices. Each microservice is concerned only with the assigned task and not the overall workflow. 

On the other hand, microservices choreography is a more bottom-up approach wherein the decision-making capability is distributed among the microservices. This makes it a decentralized system. The microservices are part of the decision/logic mechanism, and each microservice knows how and when to communicate with other services, implement tasks, and so on. 

Microservices are a part of the decision/logic mechanism, and each microservice knows how and when to communicate with other services, when to implement tasks, and so on.

Dependencies

Orchestration is characterized by actively interacting and orchestrating (for lack of a better word) events and microservices. What does this imply? It implies that while withdrawing a query request or event, one has to consider the point-to-point interconnections (dependencies) between microservices. This makes it difficult to remove or modify a defective service. 

In a choreography, however, each microservice acts autonomously, and there are no point-to-point connections. This means there is no active orchestrator that is relaying tasks to and fro. 

Organization

When there are numerous processes or events in a choreography, their execution can get messy. This is because there are too many individualistic components and no controlling unit. Things can get chaotic. 

On the other hand, in an orchestration, a controlling unit (i.e., the orchestrator) regulates the events and their execution. 

Communication

In a choreography, the services must keep track of not only their task execution at hand but also the event or the workflow of the event. The workflow represents the communication and interactions with other services. Thus, each service must be rigorous in responding to requests if they fail to execute the task. 

However, each service only carries out the assigned task in an orchestration. The orchestrator is accountable for the logic of the workflow and keeps track of the whereabouts of the microservices. 

Each service must be rigorous in responding to requests in case they fail to execute the task.

Which method should you choose?

Choosing the right communication pattern for your microservices depends on the task at hand, the scale of the organization, and a variety of other factors. Choreography is more reasonable when your application or software requires constant updates and new features. This is because any existing events or processes are not disrupted. If, however, you have a more dependent and centralized system, it can be rather challenging to perform constant updates. 

When there are complex workflows with thousands of microservices requiring levels of coordination and sync, it is better to choose an orchestration. This is because the orchestrator can act as a controlling unit and smoothly orchestrate complex workflows. 

An alternative view is adapting to a hybrid system that uses orchestration and choreography when needed. This is beneficial because it protects the system from a single point of failure (when the orchestrator fails) and allows for complex workflows. The decision depends on the context and the goal of the organization. 

I hope you now understand microservice orchestration and choreography and how they differ. 

Learn more about container orchestration and choreography

If you’d like to learn more about containers, orchestration, and choreography check out these other posts:

If your organization has an interest in choosing a microservices communication pattern, make sure to check out Architect. Architect.io provides automation and operations-based services, including microservice orchestration.


If you have questions or comments, feel free to reach out to the team! You can find us on Twitter at @architect_team.