Continuity, Transformation, and Cloud-Economics Decisions

Elasticity is not a synonym for scalability. Scalability describes the ability to add or remove capacity, while elasticity describes matching capacity to demand at the required time. An online shop with predictable annual growth can plan scaling; a sudden campaign needs elasticity that responds to the spike and releases capacity afterward. Evidence comes from measures such as response time and utilization while load changes, not merely from the presence of an Auto Scaling control. Keep this decision separate from availability: more instances in one Availability Zone may absorb demand but do not protect against failure of that Zone.

A continuity objective begins with outage impact. RTO states acceptable restoration time; RPO states the acceptable time-based amount of data loss. Distributing stateless application copies across two Availability Zones behind a load balancer addresses a data-center failure within a Region, but it is not the same as recovery from a regional disaster. Replication to another Region widens protection and adds cost, complexity, and failover testing. Multi-Region is therefore not automatically the better answer. It is justified when business impact requires it and the team can test restoration regularly.

A future audit must record GetObject and DeleteObject on one sensitive bucket and distinguish reads from writes. What action fits? Decision: Enable scoped S3 data events and inspect eventName and readOnly. Reason: Data events record the operations, and the fields distinguish reads from writes.

Instead of buying servers for a projected multi-year peak, an organization pays for resources it consumes and adjusts them with demand. Which economic principle applies? Decision: Replace upfront fixed expense with variable consumption expense. Reason: Consumption pricing shifts capacity from an upfront fixed purchase to variable expense.

A cloud provider buys and operates infrastructure against aggregated demand from many customers, achieving efficiencies that one organization generally cannot reach at the same scale. What is this concept? Decision: Economies of scale. Reason: Economies of scale result from aggregated demand, purchasing, and operations at large scale.

A small, irregular workload is eligible for an AWS-provided pay-as-you-go license. The organization owns old licenses but has not proved their cloud portability or economics. What is the sound decision? Decision: Use license included now; choose BYOL only after eligibility, vendor terms, and economics are proven. Reason: License included fits the stated flexibility, while BYOL requires proven rights, economics, and management obligations.

A migration team compares only the purchase price of an on-premises server with one cloud instance rate. What makes the TCO comparison complete? Decision: Add facilities, power, cooling, hardware lifecycle, software, and administration. Reason: TCO combines relevant direct and indirect costs instead of comparing two isolated prices.

An organization wants to justify environment-provisioning automation economically. What is the sound measure? Decision: Compare manual effort and rework before and after against automation build and run cost. Compare manual effort and rework before and after automation against its build and operating cost.

Cost decisions recur without owners, process, or organizational knowledge. Which principle fits? Decision: Build a Cloud Financial Management capability.. Cloud Financial Management builds recurring roles, knowledge, processes, and governance rather than a one-time report.

The bill fell, but successful transactions also fell. What is the sound economic measure? Decision: Measure business output against its delivery cost.. Overall efficiency relates business output to cost, so a lower bill is not success when output deteriorates.

A migration case compares only steady-state run cost and omits one-time migration work and the dual-running period. What correction is required? Decision: Include migration and dual-running costs in the business case.. A business case compares the future state with actual migration and transition-period costs.

A solution requires an additional $120,000 investment but realizes reliable revenue six months earlier; the cheaper alternative delays revenue. Which analysis determines whether speed justifies the investment? Decision: Calculate net present value or payback using incremental cost and revenue timing. A financial analysis that includes cash-flow timing shows whether earlier revenue value exceeds the incremental investment.

A Multi-AZ design provides stronger recovery but transfers 40 TB per month across zones, while another design reduces transfer with a weaker recovery boundary. What analysis is required before choosing? Decision: Model source, destination, volume, and transfer charges at different demand levels, and link each cost to the required resilience outcome. Data transfer modeling links source, destination, volume, and cost to the resilience outcome, making acceptance or redesign an informed tradeoff.

A workload was designed a year ago; demand, service prices, and managed options have since changed, but the team renews the same design automatically. Which cost-optimization practice is missing? Decision: Optimize over time through periodic review of alternatives, cost, and measured outcome. Optimize over time means periodically re-evaluating alternatives, pricing, and usage and accepting change only when the service outcome still passes.

The business estimates annual outage loss for each availability target, while each additional nine increases cost sharply. How should the target be chosen? Decision: Choose the target where reduced outage value justifies the marginal cost of the next nine. The business sets the target by comparing avoided outage value with the incremental cost of another nine.

Migration avoids 1,200 operations hours per year, but staff will be redirected to product features and no positions are eliminated. How should this benefit enter the business case? Decision: Record hours and cost as redeployed-capacity value, and claim cash savings only for an expense or contract actually removed. The business case quantifies redeployed hours and their boundary instead of automatically treating labor time as cash savings.

Three resource shapes pass the latency target but differ in error rate and cost per thousand successful transactions. What is the economic decision? Decision: Choose the lowest cost per thousand successful transactions among shapes that pass both latency and error boundaries. The decision makes unit cost and service boundaries operative rather than merely saying to measure before capacity selection.

Finance accurately allocates costs to products, but the report triggers no optimization decision. What process is missing? Decision: Assign each product an owner, review showback or chargeback regularly, and record the decision, saving or exception, and review date. The item tests post-allocation accountability and a recorded recurring decision, not merely naming CFM or producing an allocation report.

An initiative costs $180,000 in year one and produces a measurable $15,000 benefit per active team. Scenarios project 8, 12, or 20 teams. What is the break-even result? The numerical threshold is $180,000 ÷ $15,000 = 12 teams. Applying it gives $120,000, $180,000, and $300,000, showing exactly where viability changes rather than merely naming sensitivity analysis.

At a 10% annual discount rate, option A costs $100,000 now and returns $140,000 at the end of year 2. Option B costs $30,000 now and returns $50,000 at the end of each of years 1 and 2. Which option has the higher net present value? Option B wins: 50,000/1.10 + 50,000/1.10² − 30,000 ≈ $56,777, while option A is 140,000/1.10² − 100,000 ≈ $15,702. The common horizon and discount rate expose the value of timing rather than comparing nominal totals.

Successful transactions remain 1,000,000 per month. Total cost rises from $100,000 to $130,000, so unit cost rises from $0.10 to $0.13. Compute increases $5,000, storage is unchanged, and data transfer increases $25,000. Which driver explains the deterioration? With output fixed, the $30,000 cost increase fully explains the $0.03 unit-cost deterioration. Data transfer contributes $25,000, or five-sixths of that increase, so it is the dominant changed driver.

A VPN tunnel looks cheaper nominally, but alternatives differ in traffic, operating effort, availability, and outage impact. How should they be compared? Tunnel price is one component; transfer, processing, providers, operations, and failure impact can change break-even. End-to-end comparison binds cost to usage volume and availability requirements rather than a nominal price.

Archive capacity is inexpensive, but retrieval, requests, and transfer may dominate when access changes. What is the right calculation? Cost is driven by capacity and behavior; when access changes, retrieval, requests, and transfer can exceed storage. A unit model using measured mix exposes break-even instead of relying on capacity price.

Two storage options have similar per-GB prices but differ in IOPS, snapshots, operating effort, and workload success. How should the choice be made? Workload success requires performance, protection, and operation, so per-GB price alone does not represent consumed value. Comparing outcome cost after the performance gate avoids both a cheap failing option and an overbuilt one.

An organization retains cross-Region recovery copies and accepts building the recovery environment from tested definitions during an incident. Which cloud benefit is decisive? Recovery copies and tested definitions preserve recoverability, while accepting build time lets secondary compute remain off. The economic benefit is converting permanent idle capacity into consumption when needed, not claiming that RTO or testing disappears.

Adoption teams are expanding, but spend alerts have no owner or escalation path, so overruns recur. What should happen before scaling? The gap is not alert generation but accountability after the alert. Binding each threshold to an owner, escalation, and review turns the signal into action before Scale expands uncontrolled spend.

Batch workers cost $50,000 per year On-Demand. Running them on Spot costs $18,000 plus $7,000 expected interruption recompute. The noninterruptible coordinator costs $10,000 under either design. What is the economic decision? Expected cost must be compared at the same success boundaries: the mixed design is $35,000 versus $60,000, saving $25,000 while keeping the coordinator On-Demand. Nominal discount is insufficient; recompute cost and the noninterruptible component must be included.

A pricing agreement aggregates eligible usage across linked accounts, starts the benefit at 1,000 monthly units, and excludes Marketplace purchases. Three accounts record 420, 380, and 350 units, including 120 Marketplace units. What is the eligibility result? The billing relationship aggregates eligible usage, not every charge indiscriminately. Removing 120 excluded units leaves 1,030, which clears the stated threshold by 30; this is a numerical eligibility result rather than another Organizations selection.

A factory control loop must stay below 5 ms and continue through WAN loss, while the program requires on-demand cloud analytics capacity for asynchronous weekly summaries. Which operating model fits? The 5 ms and WAN-independence constraints keep control local, while asynchronous analytics can use cloud capacity. Buffering and reconciliation separate the boundaries instead of pretending a remote path can run critical control.

An organization bought servers one year ago for a nonrecoverable amount. It now compares the next three years of retaining them versus migrating, and there may be future book-value write-off, contract-exit, and disposal effects. What is the sound economic treatment? A nonrecoverable past payment should not change the ranking of future alternatives. Compare future cash flows such as operations, migration, recoverable value, contract exit, and disposal; disclose any future book-value write-off separately because it is an accounting effect, not itself a cash payment.

The organization will migrate 20% of its servers, but the facility lease, power contract, and operations team remain at the same cost until an entire hall closes. How should savings be recorded in the business case? Step-fixed facility costs do not fall proportionally with each migrated server; savings are realized when the decision actually removes a contract, hall, or operating capacity. Until then, stranded capacity and future costs remain visible, and savings plus exit or disposal effects are recorded at the exit threshold.

A Convertible Reserved Instance exchange is the operation for moving an existing commitment to eligible new attributes. The replacement must satisfy the exchange value rules, and the remaining term carries forward. Modification is narrower: it can reshape eligible size or scope details but does not substitute for an attribute change that requires exchange. Buying another discount instrument or waiting on On-Demand does not transfer the value of the existing reservation.

A Marketplace private offer targets the buyer with negotiated pricing, terms, and dimensions for acceptance. A public offer exposes standard terms and cannot embody the buyer-specific negotiation. Reserved Instance discounts apply to EC2 commitments, not third-party Marketplace legal terms.

A Standard Reserved Instance fits predictable EC2 usage when the matching attributes are stable and exchange flexibility is unnecessary. Convertible Reserved Instances trade some discount for exchange flexibility that this stable workload does not require. Compute Savings Plans are attractive when family, Region, or compute service can change; that flexibility is not the stated priority.