Book Image

Azure for Architects - Third Edition

By : Ritesh Modi, Jack Lee, Rithin Skaria
Book Image

Azure for Architects - Third Edition

By: Ritesh Modi, Jack Lee, Rithin Skaria

Overview of this book

Thanks to its support for high availability, scalability, security, performance, and disaster recovery, Azure has been widely adopted to create and deploy different types of application with ease. Updated for the latest developments, this third edition of Azure for Architects helps you get to grips with the core concepts of designing serverless architecture, including containers, Kubernetes deployments, and big data solutions. You'll learn how to architect solutions such as serverless functions, you'll discover deployment patterns for containers and Kubernetes, and you'll explore large-scale big data processing using Spark and Databricks. As you advance, you'll implement DevOps using Azure DevOps, work with intelligent solutions using Azure Cognitive Services, and integrate security, high availability, and scalability into each solution. Finally, you'll delve into Azure security concepts such as OAuth, OpenConnect, and managed identities. By the end of this book, you'll have gained the confidence to design intelligent Azure solutions based on containers and serverless functions.
Table of Contents (21 chapters)
20
Index

Cloud computing

Today, cloud computing is one of the most promising upcoming technologies, and enterprises, no matter how big or small, are adopting it as a part of their IT strategy. It is difficult these days to have any meaningful conversation about an IT strategy without including cloud computing in the overall solution discussions.

Cloud computing, or simply the cloud in layman's terms, refers to the availability of resources on the internet. These resources are made available to users on the internet as services. For example, storage is available on-demand through the internet for users to store their files, documents, and more. Here, storage is a service that is offered by a cloud provider.

A cloud provider is an enterprise or consortium of companies that provides cloud services to other enterprises and consumers. They host and manage these services on behalf of the user. They are responsible for enabling and maintaining the health of services. There are large datacenters across the globe that have been opened by cloud providers to cater to the IT demands of users.

Cloud resources consist of hosting services on on-demand infrastructures, such as computing infrastructures, networks, and storage facilities. This flavor of the cloud is known as IaaS.

The advantages of cloud computing

Cloud adoption is at an all-time high and is growing because of several advantages, such as these:

  • Pay-as-you-go model: Customers do not need to purchase hardware and software for cloud resources. There is no capital expenditure for using a cloud resource; customers simply pay for the time that they use or reserve a resource.
  • Global access: Cloud resources are available globally through the internet. Customers can access their resources on-demand from anywhere.
  • Unlimited resources: The scaling capability of cloud technology is unlimited; customers can provision as many resources as they want, without any constraints. This is also known as unlimited scalability.
  • Managed services: The cloud provider provides numerous services that are managed by them for customers. This takes away any technical and financial burden from the customer.

Why cloud computing?

To understand the need for cloud computing, we must understand the industry's perspective.

Flexibility and agility

Instead of creating a large monolithic application using a big-bang approach deployment methodology, today, applications comprise smaller services using the microservices paradigm. Microservices help to create services in an independent and autonomous manner that can be evolved in isolation without bringing the entire application down. They offer large amounts of flexibility and agility in bringing changes to production in a faster and better way. There are many microservices that come together to create an application and provide integrated solutions for customers. These microservices should be discoverable and have well-defined endpoints for integration. The number of integrations with the microservices approach is very high compared to traditional monolithic applications. These integrations add complexity in both the development and deployment of applications.

Speed, standardization, and consistency

It follows that the methodology for deployments should also undergo changes to adapt to the needs of these services, that is, frequent changes and frequent deployments. For frequent changes and deployments, it is important to use processes that help in bringing about these changes in a predictable and consistent manner. Automated agile processes should be used such that smaller changes can be deployed and tested in isolation.

Staying relevant

Finally, deployment targets should be redefined. Not only should deployment targets be easily creatable within seconds, but also the environment built should be consistent across versions, with appropriate binaries, runtimes, frameworks, and configuration. Virtual machines were used with monolithic applications but microservices need more agility, flexibility, and a more lightweight option than virtual machines. Container technology is the preferred mechanism for deployment targets for these services, and we will cover more about that later in this chapter.

Scalability

Some important tenets of using microservices are that they have an unlimited scaling capability in isolation, global high availability, disaster recovery with a near-zero recovery point, and time objectives. These qualities of microservices necessitate infrastructure that can scale in an unlimited fashion. There should not be any resource constraints. While this is the case, it is also important that an organization does not pay for resources up front when they are not utilized.

Cost-effectiveness

Paying for resources that are being consumed and using them optimally by increasing and decreasing the resource counts and capacity automatically is the fundamental tenet of cloud computing. These emerging application requirements demand the cloud as the preferred platform to scale easily, be highly available, be disaster-resistant, bring in changes easily, and achieve predictable and consistent automated deployments in a cost-effective manner.

Deployment paradigms in Azure

There are three different deployment patterns that are available in Azure; they are as follows:

  • IaaS
  • PaaS
  • SaaS

The difference between these three deployment patterns is the level of control that is exercised by customers via Azure. Figure 1.1 displays the different levels of control within each of these deployment patterns:

A tabular summary on different levels of control within each of the deployment patterns in IaaS, PaaS, and SaaS.
Figure 1.1: Cloud services—IaaS, PaaS, and SaaS

It is clear from Figure 1.1 that customers have more control when using IaaS deployments, and this level of control continually decreases as we progress from PaaS to SaaS deployments.

IaaS

IaaS is a type of deployment model that allows customers to provision their own infrastructure on Azure. Azure provides several infrastructure resources and customers can provision them on-demand. Customers are responsible for maintaining and governing their own infrastructure. Azure will ensure the maintenance of the physical infrastructure on which these virtual infrastructure resources are hosted. Under this approach, customers require active management and operations in the Azure environment.

PaaS

PaaS takes away infrastructure deployment and control from the customer. This is a higher-level abstraction compared to IaaS. In this approach, customers bring their own application, code, and data, and deploy them on the Azure-provided platform. These platforms are managed and governed by Azure and customers are solely responsible for their applications. Customers perform activities related to their application deployment only. This model provides faster and easier options for the deployment of applications compared to IaaS.

SaaS

SaaS is a higher-level abstraction compared to PaaS. In this approach, software and its services are available for customer consumption. Customers only bring their data into these services—they do not have any control over these services. Now that we have a basic understanding of service types in Azure, let's get into the details of Azure and understand it from the ground up.