Domain 1: Cloud Concepts

Domain 1 of 3. This is the smallest domain, but it builds the vocabulary for everything else, so start here.

What is cloud computing

Cloud computing is delivering computing services (servers, storage, databases, networking, software, and more) over the internet, and paying only for what you use. Instead of buying and running your own datacenter, you rent capacity from a provider like Microsoft Azure and consume it on demand.

The shared responsibility model

Security and operations are split between you and the cloud provider, and the split changes with the service type. A simple rule: you are always responsible for your data, your accounts and identities, and your devices. The provider is always responsible for the physical datacenter, the physical network, and the physical hosts.

The further you move toward SaaS, the more the provider handles and the less you do.

Cloud service types

• IaaS (Infrastructure as a Service): you rent the raw infrastructure (virtual machines, storage, networks) and manage the OS and everything above it. Most control, most responsibility. Example: an Azure Virtual Machine.
• PaaS (Platform as a Service): you get a managed platform to build and run apps without managing the OS or servers. Example: Azure App Service, Azure SQL Database.
• SaaS (Software as a Service): you use finished software over the internet and manage almost nothing. Example: Microsoft 365.

Memory aid: IaaS = you manage most, SaaS = you manage least, PaaS = in between.

Cloud deployment models

• Public cloud: services run on the provider's shared infrastructure and are offered to anyone. No CapEx, high scalability, you do not control the hardware.
• Private cloud: cloud infrastructure used by a single organization, on your own datacenter or hosted. More control and can meet strict requirements, but you carry the cost and maintenance.
• Hybrid cloud: combines public and private so workloads can move between them. Flexible: keep sensitive systems private, burst to public for scale.
• You will also hear multi-cloud (using more than one provider) and Azure Arc (manage resources across on-premises, multi-cloud, and edge from Azure).

The consumption-based model, and CapEx vs OpEx

• Capital expenditure (CapEx): a large upfront purchase of physical assets, like buying servers. On-premises is CapEx heavy.
• Operational expenditure (OpEx): paying for a service as you use it, an ongoing cost. The cloud is OpEx.
• The consumption-based model means you pay only for the resources you use, scale up or down as needed, and stop paying when you stop using. No wasted capacity, and cost becomes predictable and easier to plan.

Benefits of the cloud (know these well)

• High availability: services stay running with minimal downtime, backed by service level agreements.
• Scalability: add capacity to meet demand, either vertically (scale up, a bigger machine) or horizontally (scale out, more machines).
• Elasticity: scale automatically up and down as load changes, so you pay for what you actually need.
• Reliability: the system keeps working and recovers from failures, helped by a distributed, global design.
• Predictability: predictable performance (through scaling) and predictable cost (through the consumption model and cost tools).
• Security: a shared model plus a huge set of built in security tools and global compliance certifications.
• Governance: enforce standards and stay compliant at scale with policies and templates.
• Manageability: manage resources through the portal, command line, or code, and let the platform manage infrastructure for you.

Choosing a cloud model, by use case

• Public cloud suits startups, variable or unpredictable demand, and anyone who wants no hardware to buy and fast, cheap scale.
• Private cloud suits strict regulatory or data residency needs, or legacy systems that require full control of the environment.
• Hybrid cloud suits keeping sensitive data on-premises while bursting to the public cloud for scale, and migrating gradually rather than all at once.

Choosing a service type, by use case

• IaaS when you need to lift and shift an existing server, want full control of the operating system, or run test and development environments.
• PaaS when you want to build and ship applications quickly and focus on your code instead of servers.
• SaaS when you just need finished software such as email or collaboration, with no maintenance at all.

Cloud pricing models

• Pay as you go: pay for exactly what you use with no commitment, best for variable or short lived workloads.
• Reservations: commit to one or three years for a large discount on steady, predictable workloads.
• Spot pricing: deep discounts on spare capacity for workloads that can tolerate interruption, such as batch jobs.
• Azure Hybrid Benefit: reuse existing Windows Server and SQL Server licenses to lower cost.

Domain 1 quick recap

• Cloud computing is on demand, pay as you go computing over the internet.
• Service types: IaaS, PaaS, SaaS, in decreasing order of what you manage.
• Deployment models: public, private, hybrid.
• CapEx is upfront and physical, OpEx is ongoing and consumption based.
• In the shared responsibility model, you always own your data, identities, and devices.