Cloud Concepts: Value, Economics, and the AWS Global Model
Why the cloud changes the economics
A candidate must be able to explain the value proposition in business terms, not just list features. On-demand self-service removes procurement lead time; elasticity converts fixed capacity into variable cost; economies of scale lower unit price; and global reach lets a small team deploy near customers. A classic trap is treating "cloud" as synonymous with "virtual machines in someone else's data center." Questions often ask which benefit a scenario illustrates, so map the scenario to the specific driver: agility, cost avoidance, or resilience.
The second trap is conflating the Total Cost of Ownership (TCO) with the monthly bill. TCO includes hardware refresh, data center space and power, licensing, and staff time. A question may present a workload that looks cheaper on-premises because the invoice is small, while ignoring idle capacity and operations labor. Another recurring trap: confusing capital expense with operating expense. Migrating to AWS generally shifts spend from CapEx to OpEx, but reserved commitments can reintroduce a multi-year commitment, so read the wording carefully.
The global infrastructure model
Know the hierarchy: Regions contain Availability Zones (AZs), which are isolated data center groups with independent power and networking, connected by low-latency links. Edge locations and points of presence serve content delivery and are not places to run general workloads. A frequent question asks where to place resources for fault tolerance; the answer is usually multiple AZs within one Region, not multiple Regions, unless the requirement is geographic disaster recovery or data residency. Regions do not automatically replicate data between each other.
Worked example
A team runs a nightly batch job that takes 6 hours on a server that sits idle the other 18 hours. Rule of thumb: if average utilization is below roughly 20 percent and the workload tolerates interruption, the cloud value case is strong because you pay only for the 6 hours. If the same job must never be interrupted and runs continuously, compare reserved pricing against owned hardware TCO rather than assuming the cloud is cheaper.