Describe Cloud Concepts practice questions

Answer: A) A service credit worth a percentage of the monthly fee for the affected service

When an SLA commitment is missed, the customer usually receives a service credit, typically a percentage of the monthly fee for the affected service. Credits don't compensate for broader business losses.

• ACorrect. If the provider misses the committed service level, the SLA typically compensates the customer with service credits, usually a percentage of the affected service's monthly fee. The customer generally has to submit a claim to receive the credit.
• BIncorrect. Service credits are based on the fee for the affected service. They don't cover lost revenue, customer attrition or reputational damage.
• CIncorrect. SLAs set out credits as the compensation for missed commitments, not tier upgrades. Higher tiers may come with different SLA terms, but they aren't given out as a remedy.
• DIncorrect. The usual remedy is a service credit calculated from how far uptime fell below the commitment, not an extension of the contract term.

Answer: A) The consumption-based model, where you pay only for the resources you use

Cloud computing uses a consumption-based (OpEx) model. You pay only for what you use, so releasing resources when demand drops reduces costs, unlike up-front CapEx hardware purchases.

• ACorrect. Microsoft Learn says that in the consumption-based model you pay for the IT resources you use, and nothing more, and you release resources when you're done. Releasing resources when demand drops stops the related cost.
• BIncorrect. CapEx is up-front spending on physical infrastructure such as servers and datacenter space. That cost has already been paid whether or not the hardware is used, so turning it off doesn't stop the expense.
• CIncorrect. The shared responsibility model decides whether the customer or the cloud provider handles each management and security task. It doesn't describe how spending follows resource usage.
• DIncorrect. High availability is about keeping resources available during disruptions and is backed by SLAs. It isn't about reducing cost by releasing unused resources.

Answer: A) Their own data, user accounts, and the devices used to access the application

• AUnder SaaS, the provider manages virtually the entire underlying stack, including the finished application itself — but the customer still retains responsibility for their own data, managing their users' identities/accounts, and securing the devices used to access the service, since those inherently remain under the customer's own control.
• BRuntime and middleware are managed by the provider under SaaS, not the customer, unlike IaaS or PaaS, where more of that layer remains the customer's job.
• CThe underlying physical servers are always the provider's responsibility, across every cloud service model, never the customer's.
• DThis is incorrect — even under SaaS, the customer still retains meaningful responsibility for their own data, user accounts, and endpoint device security.

Answer: A) Private cloud

• AA private cloud is dedicated entirely to one organization, whether hosted on their own premises or by a provider on dedicated (non-shared) hardware, offering maximum control at the cost of not sharing infrastructure costs with other customers.
• BA public cloud is shared among many customers on a provider's multi-tenant infrastructure, the opposite of a dedicated, single-organization deployment.
• CA hybrid cloud combines private and public cloud resources together; this scenario describes an entirely private, standalone deployment, not a combination.
• DSaaS is a service model describing how software is delivered, not a deployment model describing where/for whom infrastructure is dedicated.

Answer: A) Capital expenditure (CapEx)

• ACapital expenditure (CapEx) refers to a large upfront investment in a physical or long-term asset (like purchased servers), which is then depreciated over its useful life on the company's books — the traditional spending pattern associated with building and owning on-premises infrastructure, in contrast to the pay-as-you-go OpEx model typical of cloud consumption.
• BOpEx refers to ongoing, recurring spending for day-to-day operations, the opposite of the large upfront asset purchase being described, which is categorized as CapEx.
• CVendor lock-in describes difficulty switching providers, an unrelated accounting concept to how a purchase is categorized.
• DElasticity describes automatic resource scaling, an unrelated accounting concept to how a purchase is categorized.

Answer: A) From capital expenditure to operational expenditure

A one-time upfront purchase is capital expenditure, and a recurring subscription for a service is operational expenditure. Moving from a perpetual license to a SaaS subscription is therefore a CapEx-to-OpEx shift.

• ACorrect. Microsoft Learn defines capital expenditure (CapEx) as upfront spending and operational expenditure (OpEx) as ongoing spending on services over time. Replacing a one-time perpetual license purchase with a recurring subscription fee for a cloud service is a move from CapEx to OpEx.
• BIncorrect. This reverses the direction. The one-time perpetual license was the upfront (CapEx) purchase, and the recurring subscription is ongoing (OpEx) spending.
• CIncorrect. The company moved away from a large one-time payment and toward recurring payments for continued use. That is closer to pay-as-you-go spending, not prepaying for reserved capacity.
• DIncorrect. Private and hybrid cloud are deployment models that describe where resources run. The scenario describes a change in how the software is paid for, not a change in deployment model.

Answer: A) Reliability

Reliability is a system's ability to recover from failures and keep functioning, and it is one of the five Azure Well-Architected Framework pillars. High availability, scalability and predictability are separate cloud benefits, and none of them is a WAF pillar.

• ACorrect. Microsoft Learn defines reliability as the ability of a system to recover from failures and continue to function. It is one of the five Well-Architected Framework pillars, along with Security, Cost Optimization, Operational Excellence and Performance Efficiency.
• BIncorrect. High availability focuses on keeping resources available as much as possible, and Azure SLAs express it as uptime guarantees. It is not one of the five Well-Architected Framework pillars. In the framework, availability is a concern covered by the Reliability pillar.
• CIncorrect. Scalability is the ability to adjust resources to meet demand, either vertically (more CPU or RAM) or horizontally (more instances). It is about handling load, not recovering from failures, and it falls under the Performance Efficiency pillar rather than being a pillar of its own.
• DIncorrect. Predictability means being able to forecast performance and cost, for example with autoscaling, load balancing and the Azure Pricing Calculator. It does not describe recovering from failures, and it is not a Well-Architected Framework pillar.

Answer: A) Multitenancy

Multitenancy means many customers (tenants) share the same physical infrastructure, with logical isolation between them. Private clouds and dedicated hosts serve a single customer, and hybrid cloud describes combining environments.

• ACorrect. Microsoft Learn describes multitenancy as shared, common infrastructure serving many tenants at the same time, with logical isolation keeping each tenant's resources and data separate. The shared infrastructure creates economies of scale.
• BIncorrect. A private cloud is used by a single organization. It isn't shared by several unrelated customers.
• CIncorrect. A hybrid cloud combines public and private cloud environments. It describes where workloads run, not several customers sharing the same physical servers.
• DIncorrect. With dedicated hosting, such as Azure Dedicated Host, a physical server is set aside for one subscription and no other customers' VMs are placed on it. That's the opposite of the shared model described.

Answer: A) Horizontal scaling (scaling out)

Adding more identical instances is horizontal scaling (scaling out). Making an existing instance bigger is vertical scaling (scaling up).

• ACorrect. Microsoft defines horizontal scaling as adding more instances of a given type of resource, so the workload is spread across several servers. Adding identical web servers behind a load balancer is exactly that.
• BIncorrect. Vertical scaling adds compute capacity to an existing resource, for example by moving it to a larger VM size. The stem says the company is not making its existing servers more powerful.
• CIncorrect. Scaling in removes instances when demand drops. It is the reverse of adding servers to handle more traffic.
• DIncorrect. Horizontal partitioning splits a dataset into separate data stores, with each holding a subset of the data. It does not add identical server instances to handle more traffic.

Answer: A) Their application code and the data it processes

• AUnder PaaS, the provider manages the OS, runtime, and middleware, letting the customer focus specifically on their own application code and its data — a smaller area of direct responsibility than under IaaS, where the customer still had to manage the guest OS and runtime themselves.
• BOS patching is handled by the provider under PaaS, not the customer — this is precisely what distinguishes PaaS from IaaS, where OS management remains the customer's job.
• CThe physical datacenter is always the provider's responsibility, in every cloud service model, never the customer's.
• DThis is incorrect — the customer still retains meaningful responsibility under PaaS, specifically for their application code and its data, even though the provider manages the underlying platform.