Question 500A utility deploys an AI route optimizer that can change live network policy with no review or rollback. After the proposed design change — Use bounded recommendations, explicit policy safety gates and reversible deployment controls. — which outcome would provide the strongest validation? A load test shows new inference replicas receive balanced traffic without exhausting transport links. A reviewed requirements register supports sizing calculations and at least two viable technology options. A test out-of-policy recommendation is rejected and the prior forwarding state can be restored. An audited inventory reconciles energy and equipment counts across branches and managed services. ExplanationCorrect Answer C:A test out-of-policy recommendation is rejected and the prior forwarding state can be restored. The result demonstrates the repair under real operating conditions. Use bounded recommendations, explicit policy safety gates and reversible deployment controls. It limits the chance that model error could reroute critical grid communications without accountability. Incorrect Answers A:Auto-scaling a workload does not improve throughput when the network bottleneck is fixed. This relates to a separate check and does not verify the specified repair. The relevant acceptance evidence is: A test out-of-policy recommendation is rejected and the prior forwarding state can be restored. B:A reviewed requirements register supports sizing calculations and at least two viable technology options. This is not a test of the affected control. AI-driven optimization requires deterministic guardrails for high-impact network changes. The required check is: A test out-of-policy recommendation is rejected and the prior forwarding state can be restored. D:Environmental reports require consistent scope and verifiable measurement assumptions. This relates to a separate check and does not verify the specified repair. The relevant acceptance evidence is: A test out-of-policy recommendation is rejected and the prior forwarding state can be restored.
Question 501A hospital protects routing adjacency with authentication but leaves device management exposed through a common user VLAN. The design team recommends: Design separate control-plane protection and management-plane access boundaries. Which architectural principle best supports that recommendation? Observability architecture must balance useful fidelity with ingestion and storage sustainability. Overlay tunnels create additional forwarding, identity and operational dependencies. Routing protocol trust and administrative access are distinct threat surfaces. End-user performance is a system property involving network and application backends. ExplanationCorrect Answer C:Routing protocol trust and administrative access are distinct threat surfaces. Applying that principle means taking the stated action. Design separate control-plane protection and management-plane access boundaries. A successful test shows: Unauthorized client VLANs cannot reach management while authenticated routing peers remain stable. Incorrect Answers A:Observability architecture must balance useful fidelity with ingestion and storage sustainability. This does not establish the proposed intervention. The governing principle here is: Routing protocol trust and administrative access are distinct threat surfaces. Apply it by: Design separate control-plane protection and management-plane access boundaries. B:Overlay tunnels create additional forwarding, identity and operational dependencies. This does not establish the proposed intervention. The governing principle here is: Routing protocol trust and administrative access are distinct threat surfaces. Apply it by: Design separate control-plane protection and management-plane access boundaries. D:End-user performance is a system property involving network and application backends. This does not establish the proposed intervention. The governing principle here is: Routing protocol trust and administrative access are distinct threat surfaces. Apply it by: Design separate control-plane protection and management-plane access boundaries.
Question 502A bank's provisioning pipeline directly changes all routers when a template variable changes, with no preflight. The design team recommends: Add schema validation, simulation and a bounded canary deployment gate before broad rollout. Which architectural principle best supports that recommendation? Continuous delivery for networks must protect high-blast-radius shared forwarding behavior. End-user performance is a system property involving network and application backends. Overlay network performance inherits underlay constraints plus extra protocol overhead. Application-aware controls may need supporting service connectivity to classify permitted flows accurately. ExplanationCorrect Answer A:Continuous delivery for networks must protect high-blast-radius shared forwarding behavior. Add schema validation, simulation and a bounded canary deployment gate before broad rollout. That is the direct response to this rule. Verify using: A malformed input fails validation and a pilot failure blocks the remaining fleet. Incorrect Answers B:End-user performance is a system property involving network and application backends. This does not establish the proposed intervention. The governing principle here is: Continuous delivery for networks must protect high-blast-radius shared forwarding behavior. Apply it by: Add schema validation, simulation and a bounded canary deployment gate before broad rollout. C:Overlay network performance inherits underlay constraints plus extra protocol overhead. This does not establish the proposed intervention. The governing principle here is: Continuous delivery for networks must protect high-blast-radius shared forwarding behavior. Apply it by: Add schema validation, simulation and a bounded canary deployment gate before broad rollout. D:Application-aware controls may need supporting service connectivity to classify permitted flows accurately. This does not establish the proposed intervention. The governing principle here is: Continuous delivery for networks must protect high-blast-radius shared forwarding behavior. Apply it by: Add schema validation, simulation and a bounded canary deployment gate before broad rollout.
Question 503A healthcare network deploys AI-based firewall policy recommendations that can become active without any audit or approval. The design team recommends: Restrict AI recommendations to reviewed candidate changes with enforceable policy safety checks. Which architectural principle best supports that recommendation? WAN overlays and segmentation do not automatically replace the security functions of branch enforcement. Data-loss protection needs coverage across actual movement paths without violating encryption constraints. Machine-generated security policy must remain subject to human governance and deterministic validation. Identity-policy resilience requires explicit cached authorization and denial semantics during partitions. ExplanationCorrect Answer C:Machine-generated security policy must remain subject to human governance and deterministic validation. The principle identifies the control actually at risk. Restrict AI recommendations to reviewed candidate changes with enforceable policy safety checks. It should produce: A synthetic overly broad rule is rejected and the rejection rationale appears in change history. Incorrect Answers A:WAN overlays and segmentation do not automatically replace the security functions of branch enforcement. This does not establish the proposed intervention. The governing principle here is: Machine-generated security policy must remain subject to human governance and deterministic validation. Apply it by: Restrict AI recommendations to reviewed candidate changes with enforceable policy safety checks. B:Data-loss protection needs coverage across actual movement paths without violating encryption constraints. This does not establish the proposed intervention. The governing principle here is: Machine-generated security policy must remain subject to human governance and deterministic validation. Apply it by: Restrict AI recommendations to reviewed candidate changes with enforceable policy safety checks. D:Identity-policy resilience requires explicit cached authorization and denial semantics during partitions. This does not establish the proposed intervention. The governing principle here is: Machine-generated security policy must remain subject to human governance and deterministic validation. Apply it by: Restrict AI recommendations to reviewed candidate changes with enforceable policy safety checks.
Question 504A manufacturing plant considers an SLA achieved because ICMP probes succeed while its industrial TCP session repeatedly resets. What is the most direct design risk if this remains unresolved? Alert storms could hide an actual outage during peak revenue hours. Production payloads larger than the tested minimum could fail unpredictably. Unaudited scripts could introduce configuration drift across campuses. Nominal reachability could conceal frequent industrial control interruptions. ExplanationCorrect Answer D:Nominal reachability could conceal frequent industrial control interruptions. The stated weakness directly creates that exposure. Service assurance must test the protocol behavior the application actually relies on. Reduce it by: Measure application handshake, session stability and request completion at the protected service boundary. Incorrect Answers A:Anomaly detection must account for predictable demand cycles and changing traffic distributions. That does not describe the principal exposure from the stated failure. The direct avoidable risk is: Nominal reachability could conceal frequent industrial control interruptions. B:End-to-end data-plane design must preserve packet characteristics through every imposed encapsulation. That does not describe the principal exposure from the stated failure. The direct avoidable risk is: Nominal reachability could conceal frequent industrial control interruptions. C:Repeatable infrastructure automation depends on reviewed source-of-truth and reproducible state. That does not describe the principal exposure from the stated failure. The direct avoidable risk is: Nominal reachability could conceal frequent industrial control interruptions.
Question 505An energy network uses anomaly models trained only on normal summer traffic and then deploys them during winter outage restoration. What is the most direct design risk if this remains unresolved? A fabricated connection could corrupt implementation plans or introduce outages. Users could be routed to an unhealthy application behind a healthy router. Model deployment could disrupt customer-facing services across the campus. Model false alarms could overwhelm operators during emergency restoration. ExplanationCorrect Answer D:Model false alarms could overwhelm operators during emergency restoration. This consequence follows from the failed control. Require representative seasonal training and governed model-drift validation before operational use. Verify improvement when: A replay of winter recovery traffic separates genuine faults from expected seasonal reroutes. Incorrect Answers A:Generated design advice is unreliable without accurate trusted network context. That does not describe the principal exposure from the stated failure. The direct avoidable risk is: Model false alarms could overwhelm operators during emergency restoration. B:Network reachability does not prove that the destination application can serve real transactions. That does not describe the principal exposure from the stated failure. The direct avoidable risk is: Model false alarms could overwhelm operators during emergency restoration. C:Model checkpoint distribution can generate bandwidth bursts far larger than inference requests. That does not describe the principal exposure from the stated failure. The direct avoidable risk is: Model false alarms could overwhelm operators during emergency restoration.
Question 506A retailer uses the same network policy for payment terminals and IoT advertising screens because both are in-store. After the proposed design change — Define payment-card trust zones with only necessary monitored exchanges to nonpayment systems. — which outcome would provide the strongest validation? New authorized pods gain only intended connectivity and stale pods lose obsolete permissions automatically. An advertising-screen test cannot initiate connections into the payment segment. A safety controller remains operational inside its limited segment and unknown devices cannot use the exception. A partner admin can perform only documented tasks while attempted global escalation is denied. ExplanationCorrect Answer B:An advertising-screen test cannot initiate connections into the payment segment. The result demonstrates the repair under real operating conditions. Define payment-card trust zones with only necessary monitored exchanges to nonpayment systems. It limits the chance that a compromised display could become a bridge into payment card infrastructure. Incorrect Answers A:New authorized pods gain only intended connectivity and stale pods lose obsolete permissions automatically. An unrelated service might produce that result while the failure continues. The relevant proof is: An advertising-screen test cannot initiate connections into the payment segment. C:Network admission controls should account for constrained devices without creating broad unauthenticated trust. This relates to a separate check and does not verify the specified repair. The relevant acceptance evidence is: An advertising-screen test cannot initiate connections into the payment segment. D:Identity federation is a high-impact security boundary requiring explicit least-privilege mapping. This relates to a separate check and does not verify the specified repair. The relevant acceptance evidence is: An advertising-screen test cannot initiate connections into the payment segment.
Question 507A media company requires cloud render jobs to ingest large raw videos but places workers far from the archive storage. After the proposed design change — Co-locate high-volume inputs with rendering compute or establish capacity-guaranteed ingestion paths. — which outcome would provide the strongest validation? Test inference sends allowed process metrics but never proprietary design files outside policy. Concurrent camera uploads leave dispatch call setup and voice quality inside emergency targets. A full-size rendering test meets job completion time with acceptable transfer spending. A customer-edge failure sends permitted cloud traffic through a surviving independent gateway. ExplanationCorrect Answer C:A full-size rendering test meets job completion time with acceptable transfer spending. Compute placement and bulk-data gravity jointly shape network performance and cost. The test establishes that the fix addresses the specific failure. Incorrect Answers A:External AI services extend the data-governance boundary to prompts and retrieved context. This relates to a separate check and does not verify the specified repair. The relevant acceptance evidence is: A full-size rendering test meets job completion time with acceptable transfer spending. B:Safety-critical real-time service should retain bounded impairment under congestion. This relates to a separate check and does not verify the specified repair. The relevant acceptance evidence is: A full-size rendering test meets job completion time with acceptable transfer spending. D:A customer-edge failure sends permitted cloud traffic through a surviving independent gateway. An unrelated service might produce that result while the failure continues. The relevant proof is: A full-size rendering test meets job completion time with acceptable transfer spending.
Question 508A telecom merger connects two identity directories and enables unrestricted administrator role mapping before trust due diligence. After the proposed design change — Federate only approved identity claims and restrict privileged roles with staged trust validation. — which outcome would provide the strongest validation? A simulated tenant compromise cannot reach another tenant's private endpoints. A room-level power failure leaves approved industrial services reachable through a protected alternate gateway. A client-group test cannot read another group's documents while approved shared tools remain accessible. A partner admin can perform only documented tasks while attempted global escalation is denied. ExplanationCorrect Answer D:A partner admin can perform only documented tasks while attempted global escalation is denied. This is the required observable result. Federate only approved identity claims and restrict privileged roles with staged trust validation. The original risk was: Overbroad federation could let compromised partner identities control the parent network. Incorrect Answers A:Tenant security requires independent network boundaries in addition to application authorization. This relates to a separate check and does not verify the specified repair. The relevant acceptance evidence is: A partner admin can perform only documented tasks while attempted global escalation is denied. B:A room-level power failure leaves approved industrial services reachable through a protected alternate gateway. This is not a test of the affected control. Identity federation is a high-impact security boundary requiring explicit least-privilege mapping. The required check is: A partner admin can perform only documented tasks while attempted global escalation is denied. C:Internal trust boundaries may be necessary for contractual confidentiality even within one employer. This relates to a separate check and does not verify the specified repair. The relevant acceptance evidence is: A partner admin can perform only documented tasks while attempted global escalation is denied.
Question 509A regional ISP aggregates many site routes behind one summary and has no mechanism to isolate a partially failed downstream island. What is the most direct design risk if this remains unresolved? A primary-edge failure could isolate stores despite healthy alternate transport. A fabricated connection could corrupt implementation plans or introduce outages. A path hash change could silently break thousands of active sessions. Traffic to the failed island could loop repeatedly within the provider core. ExplanationCorrect Answer D:Traffic to the failed island could loop repeatedly within the provider core. Use scoped more-specific announcements or health-aware withdrawal within the aggregation design. This limits the consequence because summarization must preserve enough failure signaling for usable downstream reachability. Incorrect Answers A:External route resilience requires predictable return paths and convergence behavior through each surviving exit. That does not describe the principal exposure from the stated failure. The direct avoidable risk is: Traffic to the failed island could loop repeatedly within the provider core. B:Generated design advice is unreliable without accurate trusted network context. That does not describe the principal exposure from the stated failure. The direct avoidable risk is: Traffic to the failed island could loop repeatedly within the provider core. C:ECMP across a stateful service boundary must preserve the expected session inspection context. That does not describe the principal exposure from the stated failure. The direct avoidable risk is: Traffic to the failed island could loop repeatedly within the provider core.
Question 510A telecom network uses BGP MED alone to guarantee symmetrical inbound application paths across two upstream providers. Which network architecture decision best meets these requirements? Use enforceable policy and explicit return-path analysis instead of assuming external providers honor MED. Model oversubscription and separate or schedule bulk storage traffic to maintain service priorities. Design adequate east-west fabric bisection and ECN congestion treatment for collective traffic. Size capacity and choose alternative pathing based on per-flow distribution, not just aggregate port totals. ExplanationCorrect Answer A:Use enforceable policy and explicit return-path analysis instead of assuming external providers honor MED. Inbound interdomain routing cannot be determined solely by locally preferred attributes. The practical confirmation is: Carrier tests document actual ingress for critical prefixes and stable stateful inspection. Incorrect Answers B:Fabric bisection capacity and queue design must account for simultaneous east-west demands. That is relevant to a different intervention. For the stated situation, the required action is: Use enforceable policy and explicit return-path analysis instead of assuming external providers honor MED. The governing reason is: Inbound interdomain routing cannot be determined solely by locally preferred attributes. C:Distributed training places different all-to-all throughput demands on the network than ordinary web services. That is relevant to a different intervention. For the stated situation, the required action is: Use enforceable policy and explicit return-path analysis instead of assuming external providers honor MED. The governing reason is: Inbound interdomain routing cannot be determined solely by locally preferred attributes. D:Equal-cost aggregate bandwidth does not guarantee a single large flow uses all member links. That is relevant to a different intervention. For the stated situation, the required action is: Use enforceable policy and explicit return-path analysis instead of assuming external providers honor MED. The governing reason is: Inbound interdomain routing cannot be determined solely by locally preferred attributes.