Cloud native architecture (02) - what is cloud native

Learning reference materials ：《 Enterprise Cloud native architecture ： technology 、 Service and practice ）》

01 Cloud native Overview

1.1 Yun Yuansheng defines

Wikipedia is right Cloud native （Cloud Native） The definition of ：

• Cloud native is a software development method , It makes full use of Cloud Computing , Deploy applications as microservices using the open source software technology stack . Usually , Cloud native applications are built in Docker A set of microservices running in a container , stay Kubernetes To arrange , And use DevOps2 and Git Ops Workflow management and deployment . Use Docker The advantage of container is that it can package all the software and environment configurations required for execution into an executable package . Containers run in virtualized environments , This isolates the included application from its environment .

The initial stage of going to the cloud , Most enterprises just change the application from the traditional IDC Move the computer room to the virtual machine on the cloud , This is the cloud hosting mode , Or infrastructure as a service （Infrastructure as a Service, IaaS） On the cloud .

But really make good use of the cloud , Not just infrastructure and platform upgrades , The application also needs to abandon the traditional design method , From architecture design 、 The whole software life cycle management from development mode to deployment and maintenance , It should be reconstructed based on the characteristics of the cloud , To build “ Primary cloud ” And the application of design , In this way, we can use the best architecture on the cloud 、 The best mode of operation , And make full use of the flexibility and distributed advantages of the cloud platform .

Adopt cloud native technology and management methods , We can better grow our business in “ cloud ” Or move to a cloud platform , To enjoy “ cloud ” High efficiency and continuous service ability .

1.2 Cloud native coverage

Matt Stine（《 Migrate to cloud native application architecture 》 author ） Think of cloud primitives as a collection of ideas and best practices , Including agile infrastructure （agle infrastructure）、 Microservices （microservice）、DevOps、 Continuous delivery （continuous delivery） etc. , It mainly covers the following important contents :

• Twelve elements app ： A collection of cloud native application architecture patterns ;
• Microservices ： Independently deployed Services , Each service does only one thing ;
• Agile infrastructure ： Fast 、 A platform that can repeatedly and consistently provide application environment and background services ;
• Based on application program interface （Application Programming Interface, API） Cooperation ： Release and versioning API, Allow interaction between services in the cloud native application architecture ;
• Compression resistance ： The system has good robustness , It can resist the unexpected flow impact of the outside world .

---

In a narrow sense , Cloud natives contain containers 、 Microservices 、 Continuous integration and continuous release (Continuous Integration/Continuous Delivery,CI/CD） Cloud native technology represented by , Use a new way to build 、 Deploy 、 Operation and maintenance application .

---

In a broad sense , Cloud primitives design and develop application systems completely based on distributed Cloud Architecture , It is a building software that fully uses cloud services . With the continuous development and enrichment of cloud computing technology , Many users use the cloud , It is no longer simply renting cloud vendor infrastructure in the early days （ Calculation 、 Storage 、 The Internet ） etc. IaaS resources .

02 Why cloud primitives are needed

The core idea of the cloud is elastic , In the past, we often used virtualization as the cloud platform and the interface to interact with customer systems , It not only brings flexibility to the enterprise, but also brings a certain degree of management complexity . The emergence of containers , It encapsulates a layer of virtualization , It has gradually become one of the new interfaces for cloud platform and interaction with customer systems , Application building 、 Distribution and delivery To achieve standardization at this level , A big reduction in business IT Implementation and O & M costs , Improve the efficiency of business innovation .

① Complex business requires cloud native services ： In the microservices architecture , The number of applications has increased significantly , performance 、 Consistency and other problems are becoming more and more serious , The architecture is getting more and more complex , The governance of a large number of services has also become challenging , How to deal with and solve these problems ？ The emergence of cloud native technology decouples many complexities ：

• With Docker As the representative of container technology, it realizes the decoupling between application and running environment , Many business application loads can be containerized , And application containerization satisfies agile 、 Moveable 、 The need for Standardization .
• Kubernetes The emergence of realizes the decoupling of resource scheduling and underlying infrastructure , Application and resource management and control also began to be handy , Container choreography implements resource choreography 、 Efficient scheduling .
• With Istio Service grid technology, represented by, realizes the decoupling between service implementation and service governance .

---

② Business innovation requires cloud nativity ： Internet companies often mention that they implement it dozens of times a day , Even hundreds of releases . Why is frequent publishing so important ？

• If you can release hundreds of times a day , Then you can recover from the wrong version almost immediately ; If you can recover from your mistakes immediately , Then you can take more risks ; If you can take more risks , Then you can do more “ insane ” The results of these tests may be your next competitive advantage .

---

③ Business uncertainty requires cloud nativity ： throughout IT Development history of application server ： The mainframe → minicomputer →x86 The server → virtual machine → Containers →Serverless, It is developing towards lightweight , This is also in line with the strategy of cloud native agile infrastructure . Lightweight means better elasticity , Application deployment time is reduced accordingly ： month （ The mainframe ）→ God （ minicomputer ）→ Hours （x86 The server ）→ minute （ virtual machine ）→ second （ Containers ）→ millisecond （Serverless）. Extreme flexibility is an effective means for enterprises to solve business uncertainty .

03 Cloud native design principles

① The principle of decentralization ： Centralization means a single point , In order to have good linear expansion ability , Distributed systems require decentralization , Avoid single point of failure . For the service capability of the system , As resources join , The performance and capacity of microservices can scale linearly ;

② Loose coupling principle

• Implementation of loose coupling ： That is, the service consumer does not need to rely on a specific implementation of the service contract , In this way, the internal changes of the service provider will not affect the consumer ）
• Loose coupling of time : A typical asynchronous message queuing system , Because there are intermediaries , Producers don't have to wait for a reply right away
• Loose coupling of position : Typically a service registry , The consumer does not need to know the specific location of the server directly , And are accessed through the registry lookup service
• Version of loose coupling : The consumer does not need to rely on a particular version of the service contract to work , This requires the service contract to provide downward compatibility as much as possible when upgrading ;

③ Failure oriented design principles ： In order to ensure the robustness of the system , A very important principle in the field of software design is , All external inputs and external dependencies are untrusted , Inter system dependent calls may fail at any time , It also includes hardware infrastructure services that may crash at any time , There may also be a bottleneck in the system capacity of back-end stateful services . In short, it can fail quickly when an exception occurs , Then quickly recover , To ensure that business is always online , You can't let the business stand still .

④ Stateless principle ： The cloud native application service design is stateless as much as possible , Make the business inherently scalable , During peak and low traffic periods , Automatic elastic capacity expansion based on cloud characteristics 、 Shrinkage capacity , Meet business needs . Statelessness means that when a service processes a request , Do not rely on anything other than the request itself , There will be no additional operations other than responding to the request . There are two main ways to change from stateful to stateless .

• State separation ： All state information of the server is uniformly stored in external independent distributed storage （ Caching 、 Message queue 、 database ）.
• Request with all status information ： Precede the status information , Enrich the input parameters of the request , Put the data to be processed into the input parameter through the upstream client as far as possible .

⑤ Invariance principle ： The premise of realizing the invariance principle is , Every service in the infrastructure 、 Components can be installed automatically 、 Deploy , No human intervention is required . All resources can be pulled up at any time 、 Release at any time , At the same time API Provide flexibility in a way 、 On demand calculation 、 Storage capacity . Each service or component will not change after installation and deployment , If you want to change , Just discard the old service or component , And redeploy a service or component . in addition , To improve usability , We should try our best to reduce the time to repair , You should know that the replacement speed is much faster than the repair speed .

⑥ Automation driven principles ： It is divided into continuous integration 、 Continuous deployment 、 Continuous delivery, etc , To ensure that the requirements are put forward 、 Design, development and testing , Then to the code fast 、 Safely deploy to production . Continuous integration is when a developer commits a change , Build it right away 、 automated testing , Ensure that business applications and services meet expectations , Thus, it can be determined whether the new code and the original code can be correctly integrated . Continuous deployment refers to the automatic submission of each change to the test environment using a fully automated process , Trigger automated test cases , After passing the test and verification, the application is safely deployed to the production environment , Get through development 、 test 、 Production and other links . Continuous delivery is the ability of software release , After continuous integration , It can be provided to environments such as pre release , Meet the conditions of the production environment .