This repository contains the results of my studies and understanding regarding the Exam AZ-900: Microsoft Azure Fundamentals.
Cloud Computing is the term used for defining on-demand availability of computer system resources.
In general, a cloud consists of the same computer system resources (hardware, infrastructure, etc.) you would need to have for hosting your computing solutions, but instead of owning each resource, you rent it from a cloud provider, usually sharing the resources with other companies.
Since the cloud provider is the owner of all the infrastructure & hardware, you don't need to buy new resources in order to scale your solution up. The cloud provider is prepared for scaling your solution. By centralizing the ownership of hardware and hosting, the cloud providers need to buy tons of resources and, because of it, are able to get better deals. This means the cloud providers have lots of resources available to be used by their customers.
The scalability also works the other way around - if you're not using your resource as much as the hardware and infrastructure can provide, you can scale your resources down, removing unnecessary costs.
The expenses for buying computing resources for hosting your computing solutions are considered as Capital Expenditure (CapEx), which means expenses related to acquiring equipment and infrastructure with the objective to improve the product, service or even the company premises. By using a cloud service, you can categorize this expense as a Operational Expenditure (OpEx), which translates to expenses a business incurs in its day to day operations, such as renting equipment or requiring a service from a third party. In terms of tax income, OpEx are usually better than CapEx, because the tax deduction is smaller.
With a pay-as-you-go model (pay for the resources you are currently using), the hosting costs usually are better than if you had your own on-premises infrastructure. It also removes you the need for an upfront huge expense of buying the necessary hardware, e.g.: instead of spending thousands beforehand for buying the necessary servers and databases, you can simply use them at the cloud and pay as you use for really small prices, such as a few cents per GB per month.
When a software is being provided as a service.
Example: Office 365
You don't own the software nor the code. You use an already developed and published software. You use it through the Internet, so you don't need to install it. You can't change or configure the software - you can only use what it is provided to you. You can't change the environment where this software is published as well.
When a platform is provided as a service.
Example: SQL Server
The idea is to have a whole framework being provided as a service, where you can manage and configure it. You're not only using creating a document or asking for a calculation. Instead of using a fully developed solution (as in SaaS), you're using a platform that allows you to create your own customized applications.
When the infrastructure is provided as a service.
Example: Virtual Machines
Infrastructure, in this case, means compute resources. The IaaS provides you the necessary infrastructure without the need of you buying the correspondent hardware for it. You can define, for example, on which OS will your application run or what are the network rules within your environment. The whole environment is provided by virtualization technology - this means you can't control the physical domain (e.g.: routers, CPU, graphic disks, etc).
On the contrary of what the name indicates, the Serverless model still is a server based model. It's named like that because it's meant to be dynamic: in this model you have the least configuration needed. The model automatically knows when it's needed to scale up or down (different from the other models, where you have the responsibility of configuring it). This behavior has some cons, one of the most important being the cost effectiveness: if the heavy workload at a specific period of time is always expected, having the traditional servers can be more cost effective than having a serverless solution.
A cloud that anyone can sign to and start using. You share the compute resources with other customers of the cloud provider.
Example: you sign up for Microsoft Azure and deploy your application to their cloud; the machine that hosts your application also hosts other Microsoft customers' applications.
A cloud where you're the compute resources that hosts your applications are totally owned by your company or at least only leased to your company.
Example: you bought your own servers to host your applications or you lease a server to be used only by you in Azure, in other words, no other Microsoft customer has any application hosted in this server except you.
A combination of both clouds, where you can define a part of your compute resources to be private and use the other part in the public cloud.
Example: you can have your own on-premises server that hosts your database (private cloud) and host your application on the Azure servers where other Microsoft clients also have applications being hosted (public cloud).
Regions are the way Microsoft Azure organize geographical locations in the world where their servers exist. There are currently over 60 regions. You can deploy your resource in multiple regions to, for example, have a backup if a region goes down or to improve your resource usability.
Example: if you only have customers in the US, you can use one of the US regions. If you also have customers in Spain, the Spanish customers that access your resource are going to have latency since the information needs to travel from US to Spain. It may be better to also deploy your resource in Europe as well, so the customers can get the resource from a closer place. If the European region goes down, the customers will still be able to access your resource from US, serving as a backup for them.
Some regions have their own restrictions, so you may not be allowed to deploy your application on every Azure region.
Example: you can't deploy your resource to the Azure China region because they have their own restrictions - in order to do so, you need to have a specific agreement with the Chinese company that manages Azure China.
Example 2: you can't deploy your resources to the governmental regions unless you're an employee of that governmental organization. If you are their employee, on the other hand, you can only have access to the governmental regions.
Each region has another region that is treated as its pair. This is done so you can deploy your resource in two regions in case of a backup is needed (e.g.: the resource in one region is down, so the resource in the second region can be served as a temporary backup for it).
It's almost always found in the same geographic place, but it may happen that there's no other region to be paired with, so it it's paired with some other region that it's may not be near the same geographic place. It can also happen that one region is a pair of one or more other regions.
Example: two European regions are the North Europe and West Europe, and they're a pair
Example 2: Brazil only has one region, Brazil South, so its pair is not located near Brazil but it's in US instead, the South Central US
Example 3: both East US 2 and Brazil South are paired with South Central US, but note that each combination is a separate pair - this means that East US 2 is not paired with Brazil South, only with South Central US
A group of zones inside a region. Availability zones are usually located on the same property but different buildings, having their own power, heating & cooling systems and runs on their own network. Therefore, they are completely separated physically.
Having your resources published in multiple availability zones improves resource availability since each availability zone works as a backup in case one goes down. Using multiple availability zones also prevents getting your resources down whenever Azure is updating or fixing its availability zones.
Example: you published a web app in the US East region at the availability zones 1 and 2. If the availability zone 1 goes down because of a power outage, your app will still be available because the availability zone 2 is still up.
Example 2: Azure needs to update their products (e.g.: hardware). They will only update one availability zone per time so your application goes down for the minimum needed time/doesn't go down at all because the other availability zone is still running.
If you don't choose an availability zone for your resource, Azure will choose one of them for you.
The Azure management hierarchy can be defined as the image below:
All of the resources are managed by the Azure Resource Manager (ARM), which can be accessed by three different ways:
- On browser, via the Azure Portal
- On your command line, via the Azure Command Line or Azure Power Shell
- On your Visual Studio IDE, as a REST client
The next topics will explain each of the levels of the hierarchy in the ascendant order (from the bottom to the top). Note that the lower levels inherits settings from the higher levels.
Resources are all manageable items available through Azure.
Example: Every SaaS, PaaS, IaaS or serverless computing that you use from Azure is a resource, such as VMs, web apps, databases, tags, subscriptions, etc.
Each one of them has its own cost, which takes in consideration the base costs plus external factors (such as pricing plan and other configurations you can set).
A container that groups resources for an Azure solution. You can define how you want to group the resources, usually done in a way that makes sense for your organization logically.
Example: you have an web application that also needs its own telemetry service and a functions app. Those three resources (the web application, the telemetry service and the functions app) can be grouped into the same resource group to make it easier to manage all of them together.
A subscription can be defined as a billing unit. It's where the bills for all the used resources are sent to. Each subscription is linked to a credit card for the billing, but the same credit card can be reused in more than one subscription. The bill for each respective subscription will be sent separately - this means that, even if you reuse the same credit card on multiple subscriptions, you're going to receive a different bill related to each subscription.
Subscriptions can also have different user roles, such as owner, global administrator, contributor or a simple user. Those roles can be used to give the correct rights to the correct users.
Management groups serve the purpose of grouping subscriptions or even other management groups for a better organization.
Example: your company has a development team and a marketing team. You can have a root management group called root management group, then inside of it there is a marketing management group, that has one subscription, and an IT management group, that has two subscriptions.
The meaning of "compute" is the computer is executing code - it can be your company's code or a third party application. This section describe the Azure resources that are responsible for executing computing tasks.
A "slice" of a physical machine shared with other customers. It acts and behaves as its own physical machine, but virtualized into its single piece, so you can have multiple virtual machines in the same physical machine. It's considered an IaaS.
You can configure it according to your needs, such as defining the number of CPU cores, CPU speed, RAM size and so on, such as in a physical machine.
A paradigm for running code in the cloud. In other words, you provide your code to the cloud, which will then run it. You can't change the hardware, but you can choose the plan, which will impact the performance and the costs. It's considered a PaaS.
Some examples are the application