ISC · Information Systems and Data Management · 6 practice questions
RPO vs. RTO: What Governs Database Replication?
When both replication modes are feasible, acceptable committed-data loss governs the choice, not allowable downtime. Sort one retailer’s facts: both modes are affordable and feasible, committed orders cannot be lost, and downtime cannot exceed 30 minutes.
Try one first
Hint
Which requirement, if unmet, could legally prevent HealthData from operating in a country rather than merely affecting cost or performance?
Answer B. Data residency and regulatory requirements are binding legal constraints: failing to meet them can prevent HealthData from operating in a jurisdiction. Infrastructure selection must first satisfy those residency/privacy rules (for example, using certified in‑country regions or local administrative controls where required). Only after compliance is ensured should the company optimize for latency, cost, or scalability within the compliant architecture.
Why not A: Tempting because low latency directly affects clinical effectiveness and patient safety, so it often drives architecture toward edge or local deployments. Wrong as the primary governing factor because latency solutions are only acceptable if they also meet legal/residency controls; latency cannot justify selecting an architecture that violates jurisdictional regulations.
Why not C: Tempting because minimizing capital expenditures often favors public cloud and is a strong commercial driver. Wrong as the primary factor because cost preferences cannot override statutory residency/privacy obligations that may require on‑premises systems or constrained in‑country cloud deployments, which could increase capital or operating costs.
Why not D: Tempting because scalability and speed of deployment argue for cloud-based approaches when entering markets quickly. Wrong as the governing factor because HealthData cannot lawfully deploy into a market if the chosen infrastructure fails to meet that market's legal or data‑residency requirements; scalability must be planned within compliance constraints.
Sort it
Does this fact concern whether committed orders survive a failure?
Does this fact concern when customer service or processing must resume?
Does this fact make either design possible without deciding whether its data-loss protection is sufficient?
| Item | Goes to |
|---|---|
| The database is the only source of committed order data that cannot be recreated. | RPO: Data-loss requirementRebuilding other components cannot recover committed orders missing from the surviving database copy. |
| Order processing must resume within 30 minutes after the primary site fails. | RTO: Downtime requirementThis deadline limits service interruption, not the number of committed orders that may disappear. |
| Management has approved enough budget for either synchronous or asynchronous replication. | Feasibility already establishedBudget does not distinguish the options here. The selected design must still prevent committed-order loss. |
| Database failover and traffic routing must finish within the retailer’s 30-minute recovery window. | RTO: Downtime requirementThese activities affect when users regain service. Completing them quickly does not establish that every committed order survived. |
| The retailer accepts no loss of committed orders when comparing feasible replication modes. | RPO: Data-loss requirementA zero-data-loss RPO requires synchronous replication or an equivalent guarantee against loss of committed orders. |
| Network capacity and distance support either replication mode. | Feasibility already establishedTechnical feasibility is already satisfied. It does not make asynchronous replication sufficient for the retailer’s zero-data-loss requirement. |
| Stateless application servers must be rebuilt and serving customers within the 30-minute outage allowance. | RTO: Downtime requirementServer rebuilding affects restoration time. It does not replace committed orders missing from the database. |
| The alternate site already has enough compute capacity to run the application. | Feasibility already establishedAdequate compute capacity does not establish that the standby database contains every committed order. |
| An asynchronous standby may be missing committed orders when the primary site fails. | RPO: Data-loss requirementOrdinary asynchronous replication therefore does not guarantee this retailer’s zero-data-loss RPO. |
| The 30-minute outage allowance permits delayed service restoration, not 30 minutes of lost orders. | RTO: Downtime requirementRTO measures unavailable service. The separate zero-data-loss RPO still governs replication. |
Key points
- Remember: permission to stop processing is not permission to lose orders.
- Ask what the standby could be missing, not just how quickly it can start.
- A feasible, affordable design still fails if it cannot meet the required RPO.
How the exam traps you
- Treating the 30-minute outage allowance as an acceptable replication lag. The 30 minutes measures downtime, not permitted loss of committed orders.
- Assuming quickly rebuilt application servers protect committed database transactions. Rebuilding stateless servers restores compute, not missing committed orders.
- Choosing asynchronous replication because either design fits the budget. Affordability does not establish that asynchronous replication meets a zero-data-loss RPO.
Question 2
Hint
Focus on the operational problem the users are experiencing, and set aside factors the stem says are already adequate under either option.
Answer B. The stem identifies slow disk response as the operational problem affecting a time-sensitive application. Because both options already satisfy capacity, backup, and security requirements, the primary deciding factor should be which option provides the read/write performance needed to remove the bottleneck.
Why not A: Cost is relevant in procurement, but the question asks which factor should primarily govern the decision. A cheaper option that fails to resolve the performance problem would not meet the business requirement.
Why not C: Capacity is a common focus for storage choices, but the stem explicitly states that either option has sufficient capacity. Since the issue is slow disk response, capacity is not the key differentiator here.
Why not D: Remaining book value is an accounting measure and a sunk-cost consideration; it should not drive a replacement decision. The choice should be driven by which option best meets current operational needs (performance).
Question 3
Hint
Determine which workload requires guaranteed, deterministic access to hardware (or has an explicit bare‑metal vendor restriction) rather than merely being performance‑sensitive or poorly configured.
Answer B. Deterministic, real-time control workloads that require exclusive access to specialized hardware are poor candidates for virtualization because hypervisor scheduling and interrupt handling introduce latency and jitter. Additionally, the vendor's explicit bare-metal support restriction removes virtualization as an option. Together these factors make consolidation and the expected utilization/cost benefits unlikely.
Why not A: This choice is tempting if you worry about consolidation risk, but low-utilization, independent application servers are classic virtualization candidates; consolidating them onto hypervisors typically improves utilization and reduces hardware costs.
Why not C: High random I/O can cause problems in a virtualized environment, so this distractor feels plausible; however, the issue described is an implementation/configuration omission rather than an inherent, non-remediable property of the workload. With proper paravirtualized drivers, passthrough, or storage-design adjustments the database can often be virtualized successfully.
Why not D: Dev/test needs like snapshots, cloning, and rapid provisioning are actually strengths of virtualization, so such environments typically benefit from migration rather than being exceptions to expected utilization/cost improvements.
Question 4
Hint
Focus on what separates hot, warm, and cold sites in practice: how fast they can restore operations and how much recent data they preserve.
Answer D. The primary driver of disaster recovery site selection is the business requirement for how quickly systems must be restored (recovery time objective, RTO) and how much data loss is acceptable (recovery point objective, RPO). Hot, warm, and cold sites differ mainly in speed of recovery and currency of data; select the site type that can meet the stated RTO and RPO.
Why not A: Vendor standardization can simplify operations, but it is a secondary preference. It won't change whether a site can deliver the needed RTO and RPO.
Why not B: Cost is an important constraint and hot sites are typically more expensive than warm or cold sites. However, budget should come after requirements: choosing a cheaper site that cannot meet the RTO/RPO would fail the business continuity need.
Why not C: Distance is relevant to avoid region-wide disasters and should be considered, but it does not by itself determine whether a hot, warm, or cold site can meet recovery time or data-loss targets.
Question 5
Hint
Identify which stated requirement would be hardest for management to compromise, especially since the company is willing to pay more if necessary.
Answer D. Compliance, data-location, and regulator-access requirements are nonnegotiable constraints for a bank handling nonpublic financial data. Because management is willing to pay more and the internal team needs direct control over security settings, satisfying compliance and maintaining control should govern the infrastructure choice.
Why not A: Scalability is important for many applications, but the stem emphasizes regulatory access, data-location, and control over security as the primary constraints. Those requirements should take priority over scalability when choosing infrastructure.
Why not B: Minimizing upfront capital spending is attractive and often favors cloud models, but the stem explicitly says management will accept higher cost if needed. Regulatory and control requirements outweigh capital-spending considerations here.
Why not C: Outsourcing maintenance can reduce internal workload, but the bank requires direct control over critical security configurations and must meet regulator-access and data-location rules, making maintenance convenience a secondary concern.
Question 6
Hint
Prioritize validating compliance controls and measuring real-world application performance (vendor due diligence and a pilot) before approving a migration or long-term contract.
Answer C. Because the ERP handles PCI cardholder data and is latency-sensitive, vendor claims alone are insufficient. A formal risk assessment and vendor due diligence combined with a limited pilot will validate controls, data residency, encryption, and real-world performance under expected load. Requiring appropriate SLAs and contractual controls before committing reduces the risk of outages, noncompliance, and costly premature commitments.
Why not A: This is tempting because it promises immediate cost savings and the vendor asserts compliance. It fails because vendor assertions do not substitute for due diligence; migrating without validating the provider's PCI controls, contractual SLAs, encryption, and application performance risks noncompliance and operational outages under the shared-responsibility model.
Why not B: This option is attractive because it preserves control of the environment and avoids cloud transition risks in the short term. It is problematic because it's a capital-intensive stopgap that defers an evidence-based evaluation of cloud options, scalability, and long-term TCO and does not validate vendor controls or cloud performance.
Why not D: This choice tempts by implying that single-tenant hosting automatically resolves multi-tenancy and compliance concerns. It fails because committing to a long-term contract and migrating without vendor due diligence and a pilot risks hidden control gaps, insufficient performance, or vendor lock-in; private hosting still requires verification of controls, SLAs, and real-world performance.
Common questions
Does RPO or RTO determine synchronous versus asynchronous replication?
When both designs are feasible and other requirements are satisfied, RPO governs whether asynchronous replication is sufficient. It measures acceptable loss of committed data.
Can a system allow 30 minutes of downtime but require zero data loss?
Yes. This retailer can stop processing for 30 minutes, but every committed order must survive. The outage allowance does not relax the data-loss requirement.
Can asynchronous replication guarantee a zero-data-loss RPO?
Ordinary asynchronous replication cannot guarantee it because the standby may lack committed orders at failure. This retailer needs synchronous replication or an equivalent guarantee against committed-data loss.
Practice ISC like the real exam
The free ChatCPA simulator: real exam layout, timed testlets, starting with a question on this topic. No account needed to start.
Open the free simulator →More on IT infrastructure
Questions from the ChatCPA bank of 17,658 CPA exam questions, each with a written reason for every wrong answer. ChatCPA is built by Nicholas Miller, CPA (Oregon #14907). How these pages are made. Spot an error? Tell us.