Cloud Concepts

Answer: D. Buying hardware ahead of projected demand is a capital expense committed years before the capacity is used, and idle servers are sunk cost. Paying only for capacity actually consumed converts that into a variable operating expense — one of the six advantages of cloud computing AWS names explicitly. Latency is a benefit of the global footprint, not of the CapEx-to-OpEx shift. Security controls remain a customer responsibility under the shared responsibility model. No provider guarantees workloads never fail; resilience is designed, not inherited.

Answer: D. Elasticity is the ability to acquire resources as demand rises and release them as demand falls, matching supply to load in near real time. High availability concerns surviving component failure, typically through redundancy across Availability Zones — related, but not what automatic scaling on demand describes. Durability describes the probability that stored data survives, as in the durability of S3 objects. Loose coupling is an architectural property that makes components independent; it enables scaling but is not the scaling behaviour itself.

Answer: C. Deploying into an additional AWS Region is a configuration exercise, so an application can serve a new continent in minutes instead of waiting on data centre procurement — AWS names this going global in minutes. Physical security is a benefit of the AWS facilities but is unrelated to deployment speed. Cost per request depends on architecture and usage, and is not guaranteed to fall. Nothing about a new Region redesigns an application; that remains the customer's work.

Answer: C. A highly available system aims to keep downtime very small, usually by detecting failure and recovering quickly, whereas a fault-tolerant system carries enough redundancy to continue without interruption when a component fails. They are related but not identical. Neither concept is restricted to databases. Backups support recovery but do not keep a system running through a failure.

Answer: B. Scalability describes the capacity to grow, while elasticity adds automatic expansion and contraction as load varies. They are related but not identical. Neither is restricted to a single resource type. Elasticity is precisely about capacity changing rather than staying fixed.

Answer: A. The six pillars named are exactly the pillars of the AWS Well-Architected Framework, which exists to review workloads against consistent design principles. The Cloud Adoption Framework (CAF) addresses organizational readiness for cloud migration across business and technical perspectives — it is not a workload design review. Service Catalog publishes approved provisioning products. Config records and evaluates resource configuration state, which supports a review but is not the framework itself.

Answer: A. Running workloads partly on premises and partly in the cloud, with connectivity between them, is the hybrid deployment model. Cloud-native describes applications built entirely for and in the cloud. Multi-cloud means using more than one cloud provider, which is not what a mainframe plus AWS describes. On-premises only would exclude the AWS web tier.

Answer: C. The AWS Cloud Adoption Framework organizes readiness into six perspectives covering both business and technical capability, which is exactly the list given. The Well-Architected Framework reviews individual workload designs against six pillars. Service Catalog publishes approved provisioning products. Trusted Advisor makes best-practice recommendations on an existing account.

Answer: D. Loose coupling keeps components independent, typically through a queue or an API boundary, so failure in one does not cascade into another. Vertical scaling adds capacity to a single instance and does nothing for isolation. Single-tenancy describes dedicating infrastructure to one customer. A monolith is the opposite arrangement, where components share a process and fail together.

Answer: A. A Region is a geographic area containing several Availability Zones, and each AZ comprises one or more physically separate data centres with independent power, cooling, and networking. A Region always has more than one AZ. Shared power would defeat the purpose of the separation. An AZ belongs to exactly one Region and never spans several.

Answer: B. Outposts delivers AWS-managed racks installed in the customer's own facility for workloads that must remain on premises. Local Zones place AWS infrastructure near large population centres but in AWS facilities. Edge locations cache content for CloudFront. Wavelength embeds infrastructure in telecommunications providers' 5G networks.

Answer: A. The sustainability pillar addresses reducing energy consumption and the environmental footprint of workloads. Operational excellence covers running and monitoring systems. Performance efficiency covers using resources effectively. Cost optimization covers avoiding unnecessary spend, which often correlates with sustainability but is a separate pillar.

Answer: A. Availability improves when a single failure no longer takes the service down, which requires redundancy across independent failure domains. A larger instance is still a single point of failure. Storage size affects capacity rather than uptime. Removing features changes scope without addressing failure.

Answer: B. Edge locations form a separate global network used by services such as CloudFront and Route 53 to serve users from nearby points of presence. They are distinct from Availability Zones, which host Region infrastructure. Edge locations are far more numerous than Regions. They cache and route rather than store backups.

Answer: B, D. Region choice is driven principally by latency to users and by legal requirements about where data may reside, alongside service availability and price. Region naming order is arbitrary. Edge locations are a separate global network rather than a property of a Region. Console appearance is irrelevant.

Answer: D. Rehosting moves a workload to the cloud unchanged, which is the fastest path and the usual first step in a large migration. Refactoring rewrites the application to use cloud-native services. Repurchasing replaces it with a different product, typically SaaS. Retiring decommissions a workload that is no longer needed.

Answer: C. Repurchasing moves to a different product, typically SaaS, rather than migrating the existing application. Replatforming makes targeted optimizations such as moving to a managed database while keeping the core architecture. Rehosting moves the workload unchanged. Retaining leaves the workload where it is.

Answer: A. The Cloud Adoption Framework addresses organizational readiness across six perspectives spanning business and technical capability. Designing a single workload is the role of the Well-Architected Framework. It is not a pricing model, and it is guidance rather than a service that performs migration.

Answer: A. Rehosting moves the workload unchanged, often called lift and shift. Replatforming makes targeted optimizations such as moving to a managed database. Refactoring rewrites the application. Retiring decommissions it.

Answer: C. Operational efficiency is the CAF benefit concerning faster and more effective delivery. Reduced business risk concerns resilience and security. Improved ESG performance concerns sustainability and governance outcomes. Increased revenue concerns growth.

Answer: A. Aggregated usage across many customers lets AWS achieve higher purchasing efficiency and pass lower variable costs on, which AWS lists as massive economies of scale. Elasticity is matching capacity to demand. Loose coupling is an architectural property that keeps components independent. Fault tolerance describes continued operation despite component failure. Neither elasticity, loose coupling, nor fault tolerance concerns unit pricing.

Answer: D. Reducing provisioning time from weeks to minutes lowers the cost of experimentation and lets teams iterate faster, which AWS describes as increased speed and agility. Economies of scale concern unit pricing rather than provisioning time. Durability describes the survival probability of stored data. Compliance concerns meeting regulatory obligations; faster provisioning does not by itself achieve it.

Answer: A. Provisioning on demand removes the need to forecast peak capacity years ahead, which AWS names as stopping guessing about capacity. Speed and agility concern how quickly resources can be obtained. Economies of scale concern unit pricing. Going global in minutes concerns geographic reach.

Answer: A. A managed service shifts undifferentiated operational tasks to AWS while the customer retains control of data and schema. Root access to the host is specifically not provided on RDS. Managed services are charged for. In-database users and permissions remain the customer's responsibility.

Answer: C. Because capacity is rented rather than bought, an abandoned project stops incurring cost instead of leaving purchased hardware idle, which lowers the cost of experimentation. Cloud resources can and do fail, which is why designs include redundancy. AWS does not refund spend on the basis of project outcomes. Permanence would increase rather than reduce risk.

Answer: B. Going global in minutes describes deploying to Regions worldwide quickly. Economies of scale concerns unit pricing. Provisioning speed concerns obtaining resources rather than geographic reach. Trading capital for variable expense concerns the cost model.

Answer: C. High availability means the workload survives component failure. Cost, speed, and staffing are separate concerns, although a resilient design may influence them.

Answer: D. Agility is the speed of experimentation and delivery that on-demand resources enable. Lowest cost, unlimited storage, and automatic compliance are not what agility means.

Answer: C. Operational excellence concerns running and monitoring systems and improving processes. Reliability concerns recovery from failure. Performance efficiency concerns using resources effectively. Cost optimization concerns avoiding unnecessary spend.

Answer: C. Sustainability is the pillar concerning environmental impact. The other pillars concern operations, protection, and recovery respectively.

Answer: A. Testing recovery procedures is a reliability design principle. Layered security, efficiency measurement, and global deployment belong to the other pillars.

Answer: D. Elastic capacity replaces up-front purchase, which is a core migration benefit. Latency depends on user location relative to the Region. Compliance obligations remain with the customer. Security configuration remains a customer responsibility under the shared responsibility model.

Answer: D. The People perspective addresses skills, roles, and culture. Business addresses value realisation. Governance addresses risk and control. Platform addresses the technical environment.

Answer: C. Snowball devices are shipped physically for transfers that would take too long over a network. DataSync and Transfer Acceleration optimise network transfers. Transfer Family provides managed file transfer endpoints.

Answer: A. Usage-based monthly charges are operational expense. Purchasing servers, building data centers, and buying hardware are all capital expense.

Answer: B. Managed services remove operational labour, which is a significant component of total cost of ownership. Storage and compute are still charged. Licensing depends on the engine chosen.

Answer: C. Right-sizing matches capacity to observed need. The other options all provision beyond demand.

Answer: B. Automation reduces the labour and error cost of repetition. Storage pricing, monitoring, and transfer charges are unaffected.

Answer: B. Elasticity is matching resources to demand in near real time. Durability concerns data survival, compliance concerns regulatory obligations, and interoperability concerns systems working together.

Answer: B. Purchasing servers commits spend before use, which is a fixed cost. Hourly, transfer, and per-request charges all vary with consumption.

Answer: A. Releasing resources when an experiment ends stops the cost. Experiments are charged, AWS does not refund on outcome, and the Free Tier is a separate offer.

Answer: D. Terminating idle resources removes their charge. Reserving, relocating, and snapshotting all retain or add cost.

Answer: A. Total cost of ownership includes facilities, power, cooling, network, licensing, and operations. Narrower definitions understate it.

Answer: B. Adjusting to actual demand removes the need to forecast far ahead. Some planning remains, capacity is charged, and optimization is not automatic.

Answer: B. Bring-your-own-licence and licence-included options give flexibility where terms permit. Software licensing still applies and AWS does not negotiate on the customer's behalf.

Answer: D. Managed services reduce operational labour, which is a significant cost component. They are charged, and transfer and capacity are unaffected.

Answer: D. Incremental acquisition allows starting small. There is no permanent small-company discount, the Free Tier is limited, and reservation is optional.

Answer: A. Consistency reduces errors and the cost of correcting them. Monitoring, transfer charges, and testing are unaffected.

Answer: A. Releasing capacity outside the peak removes the cost of idle time. There is no seasonal discount, automatic reservation, or transfer exemption.