Before touching a single button in the portal, it is worth understanding what problem the cloud solves and why a company decides to move its platform there. This first lesson answers three questions: what cloud computing is compared with running your own data center, what Microsoft Azure actually is and how big it really is, and what an organization gains — and loses — by adopting it. We will finish by introducing Contoso Airlines, the fictional airline that will accompany us throughout the course and whose ticket-selling platform we will build piece by piece.

This is a conceptual lesson, but it is not filler: almost every expensive mistake made in Azure (runaway bills, rigid architectures, dependencies that are impossible to undo) comes from not having understood these ideas properly at the start.

1. What cloud computing is
2. The on-premises data center and its limits
3. What Microsoft Azure is
4. The Azure catalog by service family
5. Real advantages of Azure
6. Honest trade-offs
7. Azure compared with AWS and Google Cloud
8. The Contoso Airlines case
9. Common Mistakes and Tips
10. Exercises
11. Conclusion

Cloud computing means consuming computing resources — servers, storage, databases, networking, software — over the internet and on demand, paying only for what you use, instead of buying, installing and maintaining them yourself.

The classic analogy is electricity. A nineteenth-century factory needed its own boiler and its own generator: capital tied up, maintenance staff and a fixed capacity that was excessive at night and insufficient at peak times. Today nobody generates their own electricity: you connect to the grid, consume what you need at any given moment and receive a proportional bill. The cloud does exactly that with computing capacity.

Five formally defining characteristics follow from that idea:

There are also three ways to deploy the cloud:

• Public cloud: the infrastructure belongs to the provider (Azure, AWS, Google Cloud) and is shared among customers. This is what we will cover in this course.
• Private cloud: infrastructure dedicated to a single organization, on its own premises or hosted by a third party.
• Hybrid cloud: a combination of both, with connectivity between them. It is the most common scenario in established companies, and also the one Contoso Airlines will be in during the migration.

To appreciate the cloud you have to understand what hurts in the traditional model. Contoso Airlines currently has two server rooms, one in the Barcelona office and another in the Palma office. These are their real problems:

• Buy for the peak, use at the trough. The web server was sized for November's "flight Black Friday". Eleven months a year it sits at 12% CPU, but it is paid for in full.
• Provisioning lead times. Adding two new servers means a request, a budget, an approval, an order, delivery, racking and installation: six to ten weeks. The business wants to launch a campaign in two.
• CAPEX versus OPEX. Capital is tied up in hardware that is depreciated over five years and becomes obsolete before that.
• Invisible maintenance. Patches, failing disks, UPS units, cooling, licenses, certificate renewals. Work that does not differentiate the company from its competitors.
• Theoretical disaster recovery. There is a plan in a document, but a complete failover between Barcelona and Palma has never been tested because it would have meant doubling the investment.
• Geographic reach. A customer buying from Mexico suffers the latency of crossing the Atlantic to Barcelona on every click.

The cloud does not eliminate the work, it shifts it: there are no longer disks to replace, and instead there are architectures to design, costs to govern and identities to protect. That is exactly what you will learn in this course.

Microsoft Azure is Microsoft's public cloud platform: a set of more than two hundred services deployed on a worldwide network of Microsoft-owned data centers, consumed through a common control plane (web portal, command line, API and SDKs).

Three ideas worth fixing from the very beginning:

1. Azure is real physical infrastructure. When you create a virtual machine, capacity is reserved on a specific server in a specific building. The abstraction is good, but physics still exists: that is why regions and availability zones matter (lesson 01-02).
2. Everything goes through Azure Resource Manager. It does not matter whether you use the portal, the CLI or a template: they all talk to the same management API. We will look at it in detail in lesson 01-05.
3. Services are composable building blocks. Azure is not "a product", it is a catalog. The value lies in choosing the right pieces and connecting them well.

As for scale: Azure operates in more than 60 regions spread across dozens of countries, with a presence on five continents, and it is one of the three big public clouds alongside AWS and Google Cloud. The exact figures for regions and services change every quarter, so it is best to check them in the official documentation rather than trusting a memorized number.

Seeing two hundred services in a list is overwhelming. Seeing six families is manageable. This table is the mental map of the whole course: each family corresponds, broadly speaking, to one or more modules.

A tip on method: never learn a service "because it exists". Learn the problem and then look for the service that solves it. The whole course is built that way.

• Elasticity. Capacity follows demand. Contoso can go from 2 to 20 instances of its website during the October campaign and back to 2 in November, automatically (module 2).
• Pay-as-you-go. There is no hardware purchase. You pay per hour of machine running, per GB stored or per request served. The uncomfortable corollary: whatever you leave running by mistake is also paid for.
• Global reach. Deploying in Brazil or Japan means changing a region parameter, not opening a branch office. With global delivery services, content is served close to the end user.
• Security and compliance as a foundation. Microsoft maintains certifications (ISO 27001, SOC 1/2/3, PCI DSS, ENS in Spain, among many others) covering the infrastructure, and provides encryption, identity and auditing tools. Careful: the provider's certification does not certify your application; there is a shared responsibility model that we will look at in 01-02.
• Affordable disaster recovery. Replicating data to another region is a configuration option, not a second data center.
• Integration with the Microsoft ecosystem. If the company already uses Microsoft 365, Windows Server, SQL Server or Active Directory, the continuity is real: identities federated with Microsoft Entra ID, licenses reused through Azure Hybrid Benefit (module 8) and familiar tools.
• Innovation without up-front investment. Trying out an AI service costs a few cents and an afternoon, not a year-long project.

A course that only tells you the advantages is useless for making decisions. These are the real trade-offs:

• Vendor lock-in. The more specific the service (Cosmos DB, Logic Apps, Functions), the harder it is to leave. A reasonable mitigation: use open standards where the cost of doing so is low (containers, PostgreSQL, Terraform/Bicep) and consciously accept the coupling where the benefit justifies it.
• Runaway cost if it is not governed. The ease of creating resources is also the ease of creating spend. Forgotten test environments, orphaned disks from deleted machines, oversized service tiers. That is why module 8 is a whole module, and why in lesson 01-03 we will create a budget before creating anything else.
• Learning curve. You have to learn identity, networking, cost and automation. A systems administrator who is an expert in VMware is not automatically productive in Azure on day one.
• Less fine-grained control. You do not choose the exact CPU model, and you do not walk into the server room. For workloads with very peculiar requirements (specialized hardware, licenses tied to physical machines) this can be a problem.
• Dependency on connectivity. If the office line goes down, the cloud keeps working, but you cannot reach it.
• Data sovereignty. Storing personal data of European customers involves region decisions and contracts. We cover this in 01-02.
• Prices and services change. Names, tiers and rates are revised frequently. Always check the documentation and the official calculator before committing.

A short, honest comparison, without getting into holy wars. All three are mature platforms, and Contoso's solution could be built on any of them.

Why Contoso Airlines chooses Azure: it already uses Microsoft 365 for corporate email, it has Active Directory in its offices and its booking systems run on SQL Server with current licenses. Identity continuity and license reuse tip the balance. This is not a universal truth: it is a reasoned decision in a specific context, which is exactly how these decisions should be made.

Contoso Airlines is a fictional mid-sized airline based in the Mediterranean, with around 40 aircraft and direct sales through its website. It will be our narrative thread throughout the course.

• Two server rooms: Barcelona (primary) and Palma (secondary, little more than a backup).
• The ticket-selling website runs on two physical servers behind an aging load balancer.
• The availability and pricing engine is an internal application coupled to the website.
• The booking database is on SQL Server, with a nightly copy to tape.
• Boarding passes in PDF are stored on a network file share that is already on its third full disk.
• Ground staff use a desktop application to check operations, which only works inside the corporate network.
• During peak season and one-off campaigns, the website degrades and sales are lost.

• Marta Ríos, infrastructure manager. She worries about availability, security and not losing control.
• Diego Salas, backend developer. He wants to deploy without depending on anyone and to stop configuring servers by hand.
• Nuria Peña, finance manager. She will appear in module 8 and will ask the right question: "how much is this costing us exactly, and why?".

Over the course we will assemble, piece by piece, this platform:

• Contoso Bookings: the public ticket-selling web application.
• Availability API: internal service that queries seats and prices.
• Managed booking database, with automatic backups.
• Storage for boarding passes in PDF.
• Operations dashboard for ground staff.

This diagram is deliberately high level: it does not yet say whether the website will be a virtual machine or a managed service, nor which region it will live in. Those decisions are precisely the content of the following lessons.

• Believing that "migrating to the cloud" means copying the servers as they are. Lifting the same machines into Azure (lift and shift) is a legitimate option as a first step, but if you stop there you pay for the cloud without getting its advantages. We will look at this in module 9.
• Starting to create resources before understanding billing. The correct order is: understand the model, create the account with a budget and an alert, and only then deploy. That is the order of this module, and it is no accident.
• Confusing "secure by default" with "secure". The fact that Microsoft certifies its data centers does not mean that your publicly open storage account is secure.
• Memorizing figures. The number of regions, prices and commercial names change. Memorize concepts and look up figures.
• Vocabulary tip. Use current terminology: Microsoft Entra ID (formerly Azure AD), Microsoft Defender for Cloud (formerly Security Center) and Azure AI Services (formerly Cognitive Services). You will find plenty of older documentation using the old names.
• Cost tip. We have not created anything yet, so you are not spending anything yet. From lesson 01-03 onwards you will be: every resource that is running costs real money. Get into the habit right now of switching off or deleting whatever you stop using.

List five specific problems with Contoso Airlines' current data center and indicate, for each one, which characteristic of cloud computing addresses it (self-service, elasticity, measured service, network access or resource pooling).

For each of these three Contoso decisions, write one advantage and one honest trade-off:

1. Replacing the network file share for boarding passes with object storage in Azure.
2. Using a managed database service instead of installing SQL Server on a virtual machine.
3. Choosing Azure instead of AWS.

Classify these elements of Contoso's target platform into the service families in the table in section 4: the public web application, the boarding pass PDFs, the isolation of the internal API, the booking database, the control over who accesses the operations dashboard, and the alerts for when the website responds slowly.

Solution 1 (suggested, variations accepted):

Solution 2 (suggested):

1. Object storage: advantage, practically unlimited capacity, with redundancy and a per-GB cost, and no disks to expand; trade-off, the application has to be rewritten to talk to an API instead of a network path, and public access permissions have to be watched closely.
2. Managed database: advantage, backups, patching and high availability are handled by the platform; trade-off, you lose access to the operating system and to certain advanced settings, and the cost is higher than an equivalent unmanaged VM.
3. Choosing Azure: advantage, identity continuity with the existing Active Directory and Microsoft 365, and license reuse; trade-off, greater dependency on a single provider and on its roadmap.

Solution 3:

Cloud computing replaces buying capacity with consuming capacity, and in doing so it solves the three classic pains of the on-premises data center: oversizing for the peak, taking weeks to grow and maintaining infrastructure that does not differentiate the business. Microsoft Azure is Microsoft's take on that model: an enormous catalog organized into families (compute, storage, networking, data, identity, DevOps, AI and management), deployed on a worldwide network of data centers. Its advantages — elasticity, pay-as-you-go, global reach and integration with the Microsoft ecosystem — are real, and so are its trade-offs: vendor lock-in, runaway cost without governance and a serious learning curve.

We have met Contoso Airlines, its starting point and the platform we will build together. The next question is unavoidable: when Contoso moves its website to Azure, does it put it on a virtual machine it administers itself, or on a managed service? And where on the planet does that website live? That is exactly what we will resolve in the next lesson, Service Models, Regions and Availability Zones, where we will look at IaaS, PaaS, SaaS and serverless, the shared responsibility model and how to choose where to deploy.