当前位置：网站首页>DDD application and practice of domestic hotel transactions -- Code

DDD application and practice of domestic hotel transactions -- Code

2022-07-04 14:13:00 【InfoQ】

Current explanation DDD Code landing practice content

This article 、 We will combine DDD Concept (

https://xie.infoq.cn/article/96137a34ee720466865b8bb74

Domestic hotel transactions DDD Application and Practice —— Theory Chapter ) And hotel transaction application scenarios , Focus on user interface layer 、 application layer 、 Domain layer 、 Resource layer DDD In the four tier architecture , How does the hotel transaction land DDD The implementation of , Including the implementation process 、 Standards and principles that need attention at the code level ;

One 、 Four floors are used for hotel transactions DDD Loose architecture  

DDD Four layer structure and the functions and responsibilities of each layer

User interface layer : The user interface layer is only used for processing user display and rendering of user requests and data , Should not contain domain or business logic ( Business logic is clearly defined in domain design 、 Business refers to the process that a process or behavior or an action must produce to promote the next behavior or action , Page rendering 、 The retrieval class does not belong to business logic );

application layer : Application services are located in the application layer . Application services are used to control persistent transactions and security authentication . Application services themselves do not handle business logic , Application services are lightweight , It is mainly used to coordinate the operation of domain objects , The common use of application services is ： Receive input parameters from the user interface , Then get the aggregation instance through the resource library , Then execute the corresponding command operation ;

Domain layer : Or model layer , Be responsible for expressing business concepts, business status information and business rules . This area （ Problem domain ） Complete all complex business knowledge abstractions and rule definitions ;

Infrastructure layer : One is to provide persistence mechanism for domain model ;

Layered architecture can generally be divided into strict layered architecture and loose layered architecture

Strictly layered architecture , A layer can only be coupled to the layer directly below it ;

Loose layered architecture , Then any upper layer and any lower layer are allowed to be coupled ;

Because the user interface layer and application services usually need to deal with the infrastructure , Many systems are based on a loosely layered architecture

Two 、DDD Implementation and code implementation

Brief order generation process 、 Payment process 、 Completion process 、 Reverse process

We are combining the order forward 、 Reverse business process description DDD Implementation process

1、 Building a common language for hotel transactions

2、 Four layer architecture code implementation

2.1 User interface layer code implementation example ( Order )

The code example is 、 Order agreement layer ( User interface layer ) Realization , Core logic processing , Rendering logic for receiving commands or processing user display and data ;

Because it is a back-end service 、 We put the externally provided protocol into the user interface layer implementation ;

Inject Order application layer object 、 Offer acceptance Order generation command execution 、 Order submission command execution ....

2.2 Application layer code implementation example ( Order )

The code example is 、 Order application layer implementation , Core logic , Receive input parameters from the user interface , Then get the aggregation instance through the resource library , Then execute the corresponding command operation .

principle

  Application services are located in the application layer , Used to control persistent transactions and security authentication . It does not deal with business logic , Application services are lightweight , It is mainly used to coordinate the operation of domain objects .

2.3 Domain layer code implementation example ( Order )

Order aggregation root code implementation example

Aggregation root principles and Standards

Aggregate root It's an entity , With unique identification , There's a separate life cycle , An aggregate has only one aggregate root , The aggregation root is responsible for organizing and coordinating changes to entities and value objects within the aggregation , It is not allowed to bypass the aggregation root and directly access entities in the aggregation for interaction , This will cause the aggregation to be in an inconsistent state , Violate the principle of aggregation invariance , There are certain principles to follow when designing aggregation roots ;

Aggregate root to aggregate root ： adopt ID relation ;

Aggregate the root to its internal entity , Direct object reference ;

Aggregate root to value objects , Direct object reference ;

Expose behavioral interfaces Show the expression of domain concepts , As the entrance of external communication , Only the aggregation root can be accessed externally , The aggregation root maintains the internal consistency of the aggregation ;

Domain services represent a stateless operation , It is used to implement tasks specific to a certain field . When an action does not fit on an aggregate object , You can use domain services .

Icon ： Multiple entities 、 Funding entity 、 Primary single entity 、 The product value object builds the order aggregation root 、 Combine aggregation rules and Standards 、 The business process 、 Abstract out Submit order 、 Confirm order behavior 、 Cancel the order

Include only @Getter We need to take some simple measures to prevent the packaging logo from being modified , We can setter Hide from users , Ensure that the identification will not be updated when it exists . Aggregation root creation can be handled in factory mode .

Factory code implementation example

Key points and responsibilities :

Protecting internal state is important for domain structure , Use getter&setter The party will expose the interior of the aggregation, which will cause the aggregation to be in an inconsistent state .

Large complex business systems , Entity and aggregation creation processes are complex , It's hard to create objects through simple constructors ;

The factory model solves this problem , Icon , The order aggregation root is processed in factory mode . When establishing the aggregation root , Other objects ( Entity 、 The value object ) Can automatically create , Guarantee the invariable conditions of aggregation 、 At the same time, the complexity of external packaging

Fund entity code implementation example

Substantive principles and Standards

Unique identification

Life cycle has state

Reflect relevant business behaviors

Icon ： Combined with hotel transaction capital business 、 Abstract capital entity behavior ( Room rate behavior 、 Payment behavior 、 Commission behavior, etc )、 Meet the business process ;

Entities include only @Getter We need to take some simple measures to prevent the packaging logo from being modified , We must put setter Hide from users , Ensure that the identification will not be updated when it exists .

2.4 Resource layer code implementation example

Order warehousing code implementation example

Icon ： Order warehousing layer implementation process 、 Only need to provide 2 A way

Method 1： Provide ID Get the aggregation root method

Method 2： Aggregate root unified storage 、 Ensure the internal consistency of the aggregation root ;

The repository

Our practice is that the repository serves as the isolation layer between business and data , Shield the underlying data table details , Simultaneous completion PO And DO The transformation of .

DO And PO The benefit of this transformation is that the domain layer does not rely directly on the underlying implementation , Facilitate the subsequent replacement of the underlying implementation or function migration .

Repository interfaces are defined at the domain level , The interface is implemented at the infrastructure level .

Repositories are used to hold and retrieve aggregate objects , Think of the repository as a collection of objects , Ensure the internal consistency of the aggregation root ; Not the database CRUD;

We design aggregations based on domain invariant conditions , Each request should execute a command method on an aggregate instance Aggregation design has 4 A basic principle ：

【 The principle of consistency 】

【 Keep the same conditions 】

【 A transaction modifies only one aggregate 】

【 Little aggregation 】

Consistency principle and code implementation within aggregation

Take the aggregation root as the external interface , Aggregate internal data to ensure strong consistency , The code is guaranteed by version control and simultaneous update

Consistency principle between aggregations and code implementation

Take the reverse process of hotel transactions as an example 、 User initiated application cancellation 、 Order review 、 Execute refund 、 The process involves performing aggregation 、 Order aggregation 、 Payment aggregation multiple aggregation root processing ;

How to coordinate multiple aggregation roots to ensure consistency ? First understand the standards and principles

Benchmark principle

A transaction can only update one aggregate , For example, the business process requires the system to support multiple aggregations to complete a business , So how to deal with aggregation ?

There are many technical means to achieve final consistency , May adopt MQ Mechanism , Handle changes to other aggregations in the process of event consumption .

It can meet the benchmark principle that only one aggregation can be modified for a transaction , The failure retry mechanism provided by the message queuing middleware can also be used , Complete final conformance 、 Ensure business accuracy .

Reference resources ： After the order is cancelled and approved successfully , After the approval business is completed, the approval success event will be sent , Order aggregation monitoring approval success time cancel order , After the payment aggregation listens to the cancellation event, the refund process is carried out ;

Audit domain code implementation

Step 1: execute the approval aggregation root approval cancellation

The second step is to request the audit repository to complete the aggregation root storage

Step 3 send

Audit result event

Step four The order field listens to this

Audit result event

Step five By unique identification 、 Get the order aggregation root

Step six Execute order cancellation behavior method

Step seven Request the order resource library to complete the order aggregation update

Step eight send out

Order cancellation event

Step nine Monitoring in the payment field

Order cancellation event

Step 10 Perform payment domain refund domain services 、 Complete the reverse process

Summary

The content of this article is 、 combination DDD Domain concept 、 Transaction business process 、 Explained that the hotel transaction adopts four floors DDD Implementation of architecture code 、 Apply principles and standards to each layer of architecture , Include ：