Software architecture design - software architecture style

Software architecture is a structured element with a certain form , A collection of components , Including processing components , Connecting components and data components . The processing component is responsible for processing the data , Data components are processed information , Connecting components combine and connect different parts of the architecture .
characteristic ：
1、 A pattern used to describe a particular domain of software architecture ;
2、 Software architecture provides a structure for software system 、 High level abstraction of behaviors and properties ;
3、 Software architecture is a means for project stakeholders to communicate ;
4、 Software architecture is the embodiment of early decision-making ;
5、 Software architecture is a transitive and reusable model ;
6、 Architecture restricts the quality attributes of the system , By studying software architecture, it is possible to predict the quality of software .

One 、 Software architecture modeling

Software architecture modeling is how to represent software architecture , It can be divided into five ： Structural model 、 Framework model 、 dynamic model 、 Process model and function model .
Here we mainly summarize “4+1” Model view .“4+1” View model from 5 A different perspective （ Logic 、 process 、 Physics 、 Development 、 scene ） To describe the software architecture .

Logical view ： It mainly supports the functional requirements of the system , That is, the services provided by the system to end users .
Development view ： Also known as module view , It mainly focuses on the organization and management of software modules . Software can be organized through libraries or subsystems , such , For a software system , It can be developed by different people .
Physical view ： Mainly consider how to map software to hardware , It usually takes into account the solution of the system topology 、 System installation 、 Communication, etc .
Process view ： Focus on the operating characteristics of the system , Focus on some non functional requirements , For example, the performance and availability of the system . The process view emphasizes concurrency 、 Distribution 、 System integration and fault tolerance , And the main abstract process structure in the logical view . It also defines in which thread the operations of each class in the logical view are executed . The process view can be described as a multi-level abstraction , Each level focuses on different aspects .
scene ： It can be regarded as an abstraction of those important system activities , It organically connects the four views , In a sense , Scenario is the most important requirement abstraction . When developing architecture , It can help designers find the components of the architecture and the relationship between them .

Two 、 Software architecture style

PS： The following is too long , Based on understanding , It's OK to know such a thing in practical work .
Please refer to ：《 System architecture designer tutorial 》

1、 Data flow style

The data flow style includes batch sequence and pipeline / Filter architecture style .

（1） Batch sequence architecture style . Components are a series of computing units in a fixed order , Components interact only through data transfer . Each processing step is an independent program , Each step must be started after the previous step , The data must be complete , Deliver... In a holistic manner .

（2） The Conduit / Filter architecture style . Each component has a set of inputs and outputs , The component reads the input data stream , After internal treatment , Then the output data stream is generated , After processing , Generate output data stream . This process is usually completed through the transformation of the input stream and incremental calculation , Including enriching data through calculation and adding information , By condensing and deleting refining data , Convert data by changing the recording method , Incrementally convert data, etc . Before the input is fully consumed , The output produces . Here the component is called a filter , The connector is the pipeline for data flow transmission , Pass the output of one filter to the input of another filter .

2． call / Return to style

call / The return style includes the main program / Subroutine architecture style 、 Data abstraction, object-oriented architecture style and hierarchical architecture style

（1） The main program / Subroutine architecture style . Single thread control , Divide the problem into several processing steps , Component is the main program and subroutine . Subroutines can usually be combined into modules . Procedure calls act as an interaction mechanism , That is, it acts as a connector . The calling relationship is hierarchical , Its semantic logic shows that the correctness of the subroutine depends on the correctness of the subroutine it calls .

（2） Data abstraction and object-oriented architecture style . The components of this style are objects . An object is an instance of an abstract data type . In abstract data types , The representation of data and their corresponding operations are encapsulated . The behavior of an object is reflected in its actions of accepting and requesting . Connectors are the way objects interact , Objects interact through function and procedure calls . Objects are encapsulated , Changes to one object do not affect other objects . Objects have states and operations , Also responsible for maintaining the status . This structural style includes encapsulation 、 Interaction 、 polymorphic 、 Integrate 、 Reuse and other features .

（3） Hierarchical architecture style . The hierarchical system is organized into a hierarchical structure . Component implements virtual machine in some layers . Connectors are defined by protocols that determine how layers interact , Topological constraints include constraints on the interaction between adjacent layers . The characteristic of this style is that each layer provides services for the upper layer , Use the next layer of services , You can only see the layer adjacent to yourself . The big problem is decomposed into several progressive small problems , Step by step , It hides a lot of complexity . Modify one layer , Affect up to two layers , And usually only affect the upper layer . The upper level must know the identity of the lower level , You can't adjust the order between levels . It is a common architecture design method , Can effectively simplify the design , Make the designed system structure clear , Easy to improve the ability of reuse and product maintenance . Generally speaking , The architecture of enterprise application system can be divided into The presentation layer 、 There are three layers: the middle layer and the persistence layer .

    The presentation layer . The presentation layer is mainly responsible for receiving user requests , Input to users 、 Check and control the output   system , Handle some actions of the client , Including control page Jump, etc , And present the final result information to the user . The presentation layer mainly adopts MVC Structure to achieve . The controller is responsible for receiving the user's request , And decide which model should be called to handle ; then , The model calls the middle layer to process the corresponding business logic according to the user's request , And return the data ; Last , The controller calls the corresponding view to format the data returned by the model , And present it to the user through the view .
    Middle layer . The middle layer mainly includes business logic layer components 、 Business logic layer workflow 、 Business logic layer entity and business logic layer framework . Business logic layer components are divided into two parts: interface and implementation class , Interfaces are used to define business logic components , Define the methods that business logic components must implement . Business logic components are usually designed according to modules , Each module is designed as a business logic component , And each business logic component is represented by multiple DAO Components as the basis , So as to provide the business logic service of the system externally . Business logic layer workflow can transfer documents between multiple participants according to some predefined rules 、 The process of information or tasks is automated , To achieve a desired business goal , Or promote the achievement of this goal . Business logic layer entities provide state programming access to business data and related functions , Business logic layer entity data can be constructed using data with complex architecture , This data usually comes from multiple related tables in the database , Business logic layer entity data can be used as part of business process I/O Parameter passing , The entities of the business logic layer are serializable , To maintain their current state . Business logic layer is the core component to realize system functions , In the form of containers , Facilitate the development of system functions 、 Code reuse and management .
    Persistence layer . The persistence layer is mainly responsible for the persistent storage of data , Mainly responsible for storing business data in files 、  Database and other persistent storage media . The main function of persistence layer is to provide transparent data access for business logic 、 Persistence 、 Loading capacity .

3． Independent component style

Independent component style includes process communication architecture style and event driven architecture style

（1） Process communication architecture style . Components are independent processes , Connectors are messaging . The characteristic of this style is that components are usually named , Messaging can be peer-to-peer 、 Asynchronous and synchronous mode 、 And remote procedure calls

（2） Event driven architecture style . Components do not call a procedure directly , It's about triggering or broadcasting one or more events . Processes in other components of the system are registered in one or more events , When an event is triggered , The system automatically calls all procedures registered in this event . The triggering of an event leads to the call of a procedure in another module . Components in this style are unnamed processes , The connectors that interact between them are often implicit calls between procedures （Implicit Invocation） To achieve . The main advantage of event based implicit invocation style is that it provides strong support for software reuse , It brings convenience to the maintenance and evolution of components , Its disadvantage is that the component gives up the control of system calculation .

4． Virtual machine style

Virtual machine style includes interpreter architecture style and rule-based system style

（1） Interpreter architecture style . An interpreter usually includes an interpretation engine to complete the interpretation work , A storage area containing the code to be interpreted , A data structure that records the current working state of the interpretation engine , And a data structure that records the progress of the interpreted execution of the source code . Software with interpreter style contains a virtual machine , It can simulate the execution process of hardware and some key applications . Its disadvantage is low execution efficiency .

（2） Rules based systems . Rule based systems include rule sets 、 Rule interpreter 、 The rules / Data selector and working inside
save .

5． Warehouse style

Warehouse style includes database architecture style and blackboard architecture style

（1） Database architecture style . Database architecture is the most common form of library style . There are two main categories of components , One is the central shared data source , Save the data status of the current system , The other is multiple independent processing elements , Processing elements operate on data elements .
（2） Blackboard architecture style . Blackboard architecture includes knowledge sources 、 blackboard 、 Control three parts . The knowledge source includes several different units of independent calculation , Provide problem solving knowledge , Knowledge sources respond to changes on the blackboard , Only modify the blackboard . The blackboard is a global database , Contains all the states of the solution domain , It is the only medium for the interaction of knowledge sources . Knowledge source response is controlled by the change of blackboard state . Blackboards are usually used in systems where there is no deterministic algorithm for solving problems , For example, signal processing 、 Problem planning 、 Compiler optimization and other software system design .