Microservices Use Cases

Quickly about main terms definitions.

Microservices architecture is a style of software \ application development that is based on building small independent units, where limited in size batch of services (or even a single service) defines a special process. Meanwhile, these units communicate between each other, executing the predefined logic of the application.

Developers implement it to achieve safer deployments (because each service is deployed independently and doesn't influence the work of the entire system) and decrease the large-scale user intervention.

Microservices architecture design is a kind of complicated thing to start, afterwards it serves as a mightful benefit for application maintenance and further scaling and enhancing. Moreover, microservice-based applications are a priori cloud-native, allowing implementation of the best DevOps practices.

Each company distinguishes its own benefit using such architectural style, but the most common advantages are:

• Easier deployment
  Each module is independently deployable. No need to prepare the entire application architecture to deploy.
  
• Less Dependencies
  Since all the modules are isolated, they become independent from each other. It brings clarity and "transparency" into the whole infrastructure.
  
• Faster defect identification and recovery
  Monitoring tools can quickly identify the failed module and you will be able to isolate it at once for further recovery.
  
• Less risks and fault tolerance
  The full application wouldn't crush if a module goes down. It will just continue working as usual.
  
• Easier integration with third party services
  No need to refactor the entire infrastructure to implement new tools.
  
• Polyglot system
  Microservices can be developed using almost any programming language. The most widespread are Golang, Java, Python, Ruby, NodeJS.
  
• Easier service upgrade
  If there is such a need, you can upgrade a single service instead of upgrading the entire system.

As any coin has two sides, microservice-based architectures also bear some inconveniences and tangible drawbacks. That's why microservices are not great fit for every case - though, we will talk about it later.

If microservices are built not using best practices (as it Netflix does), companies meet many challenges on the way. Though, even if the architecture is built properly, some problems may still pop up:

• Requires great efforts to implement from scratch or refactor the monolith system.
• Complicated testing across distributed environments.
• Complex infrastructure requires a lot of services what brings confusion.
• More difficult management with increase in number of services.
• Requires properly set up load balancing and message formats because of network latency and low fault tolerance.
• Sometimes can be more expensive because of remote calls, grained remote APIs, etc.

Nevertheless, these disadvantages are more easy to eliminate than disadvantages of monolithic systems.

There is no common model or unified architectural style for microservices. Anyway, there are remarkable characteristics.

1. Multiple units. Application is composed of multiple component services. Multiplication of the services is the first step towards achieving independency among services.
2. Limited functional capabilities. Each service has limited functional capabilities, i.e. so-called functional boundaries. It helps dev teams implement small tweaks or even significant changes without sync up with other teams.
3. Massive tooling. Microservices-based architectures usually involve great many tools and platform.
4. Shift from centralized systems towards decentralized ones. Dev teams try to create reproducible tools, i.e. the useful tools that can be used by other engineers to help them solve the common problems. Decentralization characteristic is also proven by the fact that each service within the microservices-based architecture utilize its own database, opposite to monoliths operating a single database across all applications.
5. Smart routing. Simple routing is the key distinguishing feature of microservices.
6. Cross-functional teams. Each team is responsible for the product, not its separate part.
7. Fault tolerance. As it was mentioned earlier, microservices have a great stability during operation. The rough description would be like the scenario below:

As it was already mentioned, not every infrastructure can be successfully rebuilt and broken into smaller parts. But industry giants showed the best microservices use cases: