· 3 min read

Separating Connectors and Service (Business Logic) - A Key Principle for Service Reusability

Separating connectors from business logic in a service-oriented architecture promotes reusability and single responsibility.

Separating connectors from business logic in a service-oriented architecture promotes reusability and single responsibility.


In a previous post, I talked about why and how we should separate Domain Layer from Technical code. Here’s another version of the same concerning service.

A Service that handles the tasks of a RestController isn’t very different from a rest controller.

We should separate the Service from the Rest Controllers because we want to be able to re-use the service.

But if Service handles rest controller responses, it isn’t very different.

An ideal service


Here’s how you separate connectors from the business logic.

  • Connectors are mediums we use to communicate with the outside world. Examples of connectors are Rest API connector (like RestController or RouterFunction+it’s Handler), and Kafka Connector.

  • Services are codes that implement business logic.

All conversions from Entity to DTO or DTO to an entity (i.e. handle the request event and provide a response in the format needed by the client) should be handled at Rest API Connector.

The Rest API Connector is responsible for calling the service method between the request and the response.

For the service method, the REST API connector is downstream. So, it does not care about what format of data it sends. It’s just responsible for executing the business logic and sending the data in whatever its format.

Example - Identifying Bad and Good Practices

Let’s examine two examples to illustrate the difference between poorly designed and well-structured code. Although the examples are based on Spring Boot, the principles apply to any project.

Bad Practice: Violation of Single Responsibility

// @Controller Code
ResponseEntity<UserResponse> postUsers(UserRequest userRequest){
    return userService.saveUser(userRequest); // 1

// @Service Code
ResponseEntity<UserResponse> saveUser(UserRequest userRequest){ // 2
    UserEntity unsaved = objectMapper.convertValue(userRequest, UserEntity.class); // 3
    unsavedUser.setUserId(sequenceService.getNextId()); // 4
    UserEntity savedUser = userRepo.save(unsaved); // 5
    UserResponse userResponse = objectMapper.convertValue(savedUser, UserResponse.class); // 6
    return ResponseEntity.ok().body(userResponse); // 7

Here is why the above code is bad:

  • 1: @Controller is just calling the @Service and returning its value. It isn’t doing anything except @PostMapping. If that’s the case what’s the purpose of a @Controller, just mapping?
  • 2: The @Service is handling receiving the UserRequest and returns a ResponseEntity of UserResponse.
    • The problem both UserRequest and ResponseEntity<UserResponse> are mapped to REST API. I can’t reuse the method save mechanism for other connections or from within other services.
    • For example, let’s say I have to save a user from another @Service. To use the same saveUser method, I’ll have to create a UserRequest object and the parse the ResponseEntity<UserResponse>.
    • This is a highly inconvenient conversion.
  • 3 to 7: Here the @Service isn’t singularly responsibly.
    • While the name of the method is saveUser, it is also handling the conversion of request to entity and entity to the response.
    • @Service isn’t just handling business logic.

Good Practice: Preserving Single Responsibility

// @Controller Code
ResponseEntity<UserResponse> postUsers(UserRequest userRequest){
    UserEntity unsavedUser = objectMapper.convertValue(userRequest, UserEntity.class); // 1
    UserEntity savedUser = userService.saveUser(unsavedUser); // 2
    UserResponse userResponse = objectMapper.convertValue(savedUser, UserResponse.class); // 3
    return ResponseEntity.ok().body(userResponse); // 4

// @Service Code
UserEntity saveUser(UserEntity unsavedUser){ // 5
    unsavedUser.setUserId(sequenceService.getNextId()); // 6
    return userRepo.save(unsaved); // 7

The above code is good because:

  • The @Controller handles the conversion from Request object to Entity and Entity to Response.
  • The @Service handles the business logic
  • The @Service only deals with entities and not request, response objects and therefore can be used at many places easily.


  • A service method should only deal with business logical.
  • The connectors like @Controller or other channels should deal with conversion from request object to response object.
  • This way the service remains open to be used in mutiple channels or in multiple other services.

Back to Blog