[AZ-900] Module 1: Cloud concepts 정리

강의 내용 + 책 내용(강의를 들을 당시 e-book을 제공해줬는데 내용 자체는 MS Doc 들어가면 있는 그 내용 그대로)

docs.microsoft.com/en-us/learn/modules/intro-to-azure-fundamentals/introduction

Introduction - Learn

 

[Module 1] Cloud concepts

Cloud

Run IT services(servers, storage, DB, networking, software, analytics, and intelligence) online without managing a HW

Offer faster innovation, flexible resources, and economies of scale

 

Why is it cheaper?

- “Pay-as-you-go pricing model”

    *  Lower operating costs

    *  Run infrastructure more efficiently

    *  Great scalability

- “Rent compute power and storage from others’ data center”

    *  Treat cloud resources

    *  Cloud provider takes care of maintaining the infrastructure

 

Why need to move?

- Teams deliver new features to their users at record speeds(업데이트가 잦음)

- Users expect an increasingly rich and immersive experience with their devices and with SW(사용자의 기대 폭발적 증가)

- 이때 cloud는…

    *  Almost limitless pool of raw compute, storage, and networking components

    *  Speech recognition and other cognitive services that help make application stand out from the crowd

    *  Analytics services that deliver telemetry data from SW and devices

 

Characteristics(5)

- On-demand: 사용하고 싶으면 가입하고 바로 사용 가능. 회사에 구매하겠다고 따로 연락을 취할 필요X

- Resource pool: Offer a lot of products

           Ex) server 필요-> compute-virtual machine, 저장소 필요-> storage, DB 필요-> database

- Elasticity: flexible. 부족하면 늘리고 너무 많으면 줄이기 가능

- Via Internet: 인터넷만 있으면 바로 이용 가능

- Pay as you go: 사용한 만큼 pay

 

Azure

- MS에서 개발한 cloud service

- Give freedom to build, manage, and deploy applications on a massive global network

 

Azure offers…

-       Be ready for the future: Continuous innovation supports development

-       Build on your terms: Users have choices. 오픈소스, 모든 언어, 프레임워크 지원 노력-> User가 원하는 곳에 원하는 방식대로 deploy 가능

-       Operate hybrid seamlessly: 원활한 하이브리드 운영(사내, 클라우드 등)

-       Trust your cloud: 보안 굿

 

Azure로 할 수 있는 것

100개 이상(강의에서 200개 이상이라고 하셨는데… 책은 100개라고 나옴… 뭐지…)의 서비스 제공

 

How does Azure work?

“Virtualization”: Hypervisor(추상화 계층)를 사용해서 컴퓨터의 HW와 OS간의 결합 분리. Hypervisor로 VM에서 실제 컴퓨터와 해당 CPU의 모든 함수 에뮬레이팅. 여러 VM 동시 실행 가능. 에러 발생 HW 용량 최적화.

각 data center에는 서버로 채워진 미니 랙이 있음. 각 서버당 여러 VM을 실행하는 Hypervisor 포함. 네트워크 스위치는 이런 모든 서버에 대한 연결 제공. 각 랙에 있는 하나의 서버는 패브릭 컨트롤러라는 SW 실행. 각 Fabric controller는 orchestrator(사용자 요청(Orchestrator Web API 사용)에 응답, Azure에서 발생하는 모든 작업 관리) 라는 다른 SW에 연결

Ex)

사용자가 VM 생성 요청->Orchestrator가 필요한 모든 것을 패키지로 만들고 최상의 서버 랙 선택-> 패키지와 request를 Fabric controller로 보냄->Fabric controller에서 VM 생성->사용자가 VM 연결 가능

 

Azure portal

Web-based, unified console. Provides an management system with GUI

 

Azure Marketplace

Solutions & services 제공

 

Products

COMPUTE: Let scale computing capability

Azure Virtual Machines	윈도우나 리눅스 VM 호스팅(수동)
Azure Virtual Machine Scale Sets	VM Scaling(자동)
Windows Virtual Desktop	Azure service 맞음… Combined with appropriate licenses, services, and resources, delivers a complete virtualized multi-user Windows 10 (or a single-user Windows 7) experience together with Office 365 ProPlus 대충 사원들이 cloud를 통해 사용할 수 있는 데스크톱 업무환경(OS랑 office SW 깔림) 제공하는 product
Azure Container(mini application: app 기능 부품화) Instances	서버나 VM 없이 run containerized apps
Azure Kubernetes Service	Cluster management for VMs that run containerized services
Azure App Service	quickly build, deploy, and scale enterprise-grade web, mobile, and API apps running on any platform
Azure Service Fabric	항상 사용할 수 있고 확장 가능한 분산형 앱을 구축하고 운영
Azure Batch	대규모 병렬 및 HPC(고성능 컴퓨팅) 일괄 작업을 Azure에서 효율적으로 실행
Azure Functions	An event-driven, serverless compute service(1 function)
Azure Logic Apps	Events-driven, serverless compute service(multiple functions)

 

 

NETWORKING: Let connect cloud and on-premise infrastructure(VPN, load balancing)

Azure Virtual Network	Connects VMs to incoming VPN connections -link resources together in your on-premises environment and within your Azure subscription -filter traffic between subnets -link virtual networks together by using virtual network peering
Azure Load Balancer	Balances inbound and outbound connections to applications or service end points
Azure Application Gateway	웹 애플리케이션에 대한 트래픽을 관리할 수 있도록 하는 웹 트래픽 부하 분산 장치
Azure VPN Gateway	공용 인터넷을 통해 Azure 가상 네트워크와 사내 간에 암호화된 트래픽을 전송하는 데 사용되는 특정 유형의 가상 네트워크 게이트웨이
Azure DNS	Microsoft Azure 인프라를 사용하여 이름 확인을 제공하는 DNS 도메인에 대한 호스팅 서비스. Azure에 도메인을 호스트하면 다른 Azure 서비스와 동일한 자격 증명, API, 도구 및 대금 청구를 사용하여 DNS 레코드를 관리
Azure Content Delivery Network	Delivers high-bandwidth content to customers globally
Azure ExpressRoute	Connects to Azure over high-bandwidth dedicated secure connections. Extend on-premises network into Azure
Azure Traffic Manager	DNS 기반 트래픽 부하 분산 장치. 글로벌 Azure 지역에서 공용 연결 애플리케이션에 트래픽을 배포할 수 있음. 고가용성과 빠른 응답성을 갖춘 퍼블릭 엔드포인트를 제공.
Azure DDoS Protection	실제 트래픽 사용률을 모니터링하고 이를 DDoS 정책에 정의된 임계값과 지속적으로 비교. 트래픽 임계값이 초과되면 DDoS 완화가 자동으로 시작됩니다. 트래픽이 임계값 아래로 돌아오면 완화가 중단
Azure Firewall	Azure Virtual Network 리소스를 보호하는 관리되는 클라우드 기반 네트워크 보안 서비스
Azure Virtual WAN	많은 네트워킹, 보안 및 라우팅 기능을 결합하여 단일 운영 인터페이스를 제공하는 네트워킹 서비스
Azure Network Watcher	Provides tools to monitor, diagnose, view metrics, and enable or disable logs for resources

 

 

STORAGE: Disk/file/blob/archival storage

Disk Storage	provides disks for Azure virtual machines
Azure Blob storage(Container Storage)	Storage service for very large objects, such as video files or bitmaps.
Azure File storage	File shares that can be accessed and managed like a file server.
Azure Queue storage	A data store for queuing and reliably delivering messages between applications.
Azure Table storage	Stores non-relational structured data (also known as structured NoSQL data)

Storage access tiers

Hot	Cool	Archive
매우 자주	그럭저럭	안 자주

Common characteristics

• Durable and highly available with redundancy and replication.
• Secure through automatic encryption and role-based access control.
• Scalable with virtually unlimited storage.
• Managed, handling maintenance and any critical problems for you.
• Accessible from anywhere in the world over HTTP or HTTPS.

 

 

MOBILE: Build and deploy cross-platform and native apps. Sent notifications…

Other features of this service include:

• Offline data synchronization.
• Connectivity to on-premises data.
• Broadcasting push notifications.
• Autoscaling to match business needs.

 

 

DB

Azure Cosmos DB	NonSQL용
Azure SQL Database	Fully managed relational database with auto-scale, integral intelligence, and robust security.
Azure Database for MySQL/PostgreSQL	MySQL/PostgreSQL용
Azure SQL Managed Instance	provides the broadest SQL Server database engine compatibility with all the benefits of a fully managed platform as a service
SQL Server on Azure Virtual Machines	Service that hosts enterprise SQL Server apps in the cloud
Azure Synapse Analytics	Fully managed data warehouse with integral security at every level of scale at no extra cost
Azure Database Migration Service	Service that migrates databases to the cloud with no application code changes
Azure Cache for Redis	Fully managed service caches frequently used and static data to reduce data and application latency
Azure Database for MariaDB	MariaDB용

 

 

WEB: Help build, deploy, manage, and scale web application

Azure App Service	Quickly create powerful cloud web-based apps.
Azure Notification Hubs	Send push notifications to any platform from any back end.
Azure API Management	Publish APIs to developers, partners, and employees securely and at scale
Azure Cognitive Search	Deploy this fully managed search as a service.
Web Apps feature of Azure App Service	Create and deploy mission-critical web apps at scale
Azure SignalR Service	Add real-time web functionalities easily

 

 

IoT: Connect, monitor, and manage all the IoT assets

Azure Sphere	Sensor 관리
Azure IoT Hub	Messaging hub that provides secure communications between and monitoring of millions of IoT devices. Edge processing
IoT Central	Fully managed global IoT software as a service (SaaS) solution that makes it easy to connect, monitor, and manage IoT assets at scale. Center cloud. Compile result.
IoT Edge	Fully managed service that allows data analysis models to be pushed directly onto IoT devices, which allows them to react quickly to state changes without needing to consult cloud-based AI models

 

 

Big Data: With large amount of data, run analytics at a massive scale, make decisions based off of complex queries

Datalake	Storage(table X)
Data Warehouse	Storage(table form)
Azure Databricks	Analysis(Apache Spark) Integrate this collaborative Apache Spark-based analytics service with other big data services in Azure
Azure HDInsight	Analysis(Other open source, ex- Hadoop) Process massive amounts of data with managed clusters of Hadoop clusters in the cloud.
Azure Synapse Analytics	Storage+Analysis Run analytics at a massive scale by using a cloud-based enterprise data warehouse that takes advantage of massively parallel processing to run complex queries quickly across petabytes of data.

 

 

AI: With existing data, forecast future behaviors. Use ML to build, train, and deploy models on cloud

Data/Train           Execute(pretrainded one: Cognitive services)

 

Cognitive services

Vision	Use image-processing algorithms to smartly identify, caption, index, and moderate your pictures and videos.
Speech	Convert spoken audio into text, use voice for verification, or add speaker recognition to your app.
Knowledge mapping	Map complex information and data to solve tasks such as intelligent recommendations and semantic search.
Bing Search	Add Bing Search APIs to your apps and harness the ability to comb billions of webpages, images, videos, and news with a single API call.
Natural Language processing	Allow your apps to process natural language with prebuilt scripts, evaluate sentiment, and learn how to recognize what users want.

Azure Machine Learning Service	Data/train+execute Cloud-based environment you can use to develop, train, test, deploy, manage, and track machine learning models. It can auto-generate a model and auto-tune it for you. It will let you start training on your local machine, and then scale out to the cloud
Azure ML Studio	Collaborative visual workspace where you can build, test, and deploy machine learning solutions by using prebuilt machine learning algorithms and data-handling modules

 

 

DevOps: Create build and release piplelines

Storage code      Build, Develop        Test

Azure Repo	Storage code
Github	Storage code
Azure pipeline	Build, Develop
Github with action	Build, Develop
Azure DevTest Labs	Test Quickly create on-demand Windows and Linux environments to test or demo applications directly from deployment pipelines.
Azure DevOps	Storage code+build, deploy+test Use development collaboration tools such as high-performance pipelines, free private Git repositories, configurable Kanban boards, and extensive automated and cloud-based load testing. Formerly known as Visual Studio Team Services.

 

Account architecture

1.     Register an account

2.     Subscription(payment method)

3.     Resource(project) group

4.     Resources

 

Cloud deployment models

- Public cloud: offered over the public internet. Available to anyone who wants to purchase. Cloud resources(server, storage etc) are owned and operated by cloud service provider

- Private cloud: consists of computing resources used exclusively by users from one business or org. can be physically located at organization or hosted by a third party service provider

- Hybrid cloud: combination of public and private cloud

 

Comparison

Public: No capital expenditures to scale up(HW를 살 필요가 없어서). Application 제공/제공 해제 신속하게 가능. 사용한만큼만 지불

Private: HW 사야함. 리소스와 보안 문제 온전하게 기업이 해결해야함. HW 유지 보수 책임 있음.

Hybrid: 가장 flexible. 기업이 어디에 application을 run 해야할지 정해야함. 기업이 보안, 규정 준수 또는 법적 요구사항 제어.

 

Cloud benefits

High availability: provide a continuous user experience with no apparent downtime, even when things go wrong. “Backup server”

Scalability: Vertically-RAM/CPU… Horizontally-Adding instances of resources such as VMs

Elasticity: Autoscaling

Agility: Template->Quick deploy

Geo-distribution: 원하는 region의 datacenter 선택 가능

Disaster recovery: safe in the event of disaster

 

High availability VS Disaster Recovery

Backup server를 같은 나라의 data center에 두었는가

                다른

 

High availability VS Fault tolerance

Service maintain | Quality&service maintain

 

Scalability VS Elasticity

Manually | Automatically

 

CapEx : Spending money on physical infrastructure

OpEx: Pay as u go

 

Cloud Service Model

IaaS(Infrastructure-as-a-Service)

Cloud provider가 HW 신경써주지만 OS나 network는 cloud tenant 책임. Ex) Virtual Machine

😊

No CapEx, Agility, Management, consumption-based model, no deep technical skills required, cloud benefits, flexibility

 

PaaS(Platform-as-a-Service)

Cloud provider가 VM, networking 관리해줌. Tenant는 application deployment에 집중 가능

Ex) DB, Web application deployment

😊

No CapEx, Agility, Management, consumption-based model, cloud benefits, Productivity

 

☹

Platform limitation-might affect how an application runs

 

SaaS(Software-as-a-Service)

다 만들어놓은거 사용만 하면 됨

Ex) MS Office 365

 

😊

No CapEx, Agility, Pay-as-you-go pricing model, no deep technical skills required, flexibility

☹

Software limitations

 

IaaS	PaaS	SaaS
The most flexible cloud service.	Focus on application development.	Pay-as-you-go pricing model.
You configure and manage the hardware for your application.	Platform management is handled by the cloud provider.	Users pay for the software they use on a subscription model.

 

Serverless computing->chap 3

 

 

시험날짜 정하려고 했더니 오류가 생겨서 이메일 보냈는데 화요일에 돌아오신다는 자동응답 메시지가... 그냥 마음 비우고 다다음주에나 볼 생각 해야겠다...