ISC CCSP Practice Test 40
Question 390
Before changing production, which focused diagnostic is most useful for identifying the true dependency? A DevOps team deploys new container versions without versioned release records and cannot identify which build introduced a privilege regression. Select the ONE best answer.
- Inspect DS and DNSKEY records, validation error logs, NTP, resolver configuration and controlled spoofed-response tests.
- Review cloud firewall rules, gateway permissions, MFA settings and attempted session event records.
- Review artifact registry history, deployment manifests, pipeline IDs, approvals and rollback rehearsal events.
- Compare image baseline, live SSH policies, change records, exception expirations and compliance scan outcomes.
Explanation
Correct Answer
C:Review artifact registry history, deployment manifests, pipeline IDs, approvals and rollback rehearsal events. This observation directly isolates the suspected dependency before a risky change. Release governance provides provenance and coordinated acceptance across application and platform changes. The appropriate intervention is: Implement controlled release and deployment management with artifact identity, signed approvals, change scope and rollback linkage.
Incorrect Answers
A:DNSSEC protects authenticity of DNS responses when its signatures and trust chain are correctly maintained and validated. The focused check for the reported condition is: Review artifact registry history, deployment manifests, pipeline IDs, approvals and rollback rehearsal events.
B:Privileged remote protocols should be protected by verified identities, limited network reachability and recorded access. The focused check for the reported condition is: Review artifact registry history, deployment manifests, pipeline IDs, approvals and rollback rehearsal events.
D:Secure baselines drift when emergency changes are not reconciled through controlled maintenance and evidence. The focused check for the reported condition is: Review artifact registry history, deployment manifests, pipeline IDs, approvals and rollback rehearsal events.
Question 391
A change engineer must select a justified corrective intervention. Which should be approved? A cloud penetration test is scheduled without confirming which managed services and shared tenant infrastructure are in the provider's permitted testing scope. Select the ONE best answer.
- Establish incident escalation paths, secure evidence requests and contractual response points for both customer and provider teams.
- Adapt forensic plans to the actual cloud service evidence boundary and prearrange provider cooperation for unavailable layers.
- Establish written scope, provider rules, testing windows, authorization and rollback before conducting adversarial validation.
- Validate supported host capabilities, nested virtualization policies and guest resource and security controls before activation.
Explanation
Correct Answer
C:Acceptance criterion: The test exercises only approved targets, records useful findings and produces no unapproved cross-tenant disruption. Penetration testing in shared cloud environments must avoid unauthorized interference with provider or other tenant assets.
Incorrect Answers
A:Here the decisive dependency is: Penetration testing in shared cloud environments must avoid unauthorized interference with provider or other tenant assets. This action belongs to a different situation: A cloud breach triggers internal containment, but the incident plan has no named contacts for the service provider to preserve provider-held logs.
B:Here the decisive dependency is: Penetration testing in shared cloud environments must avoid unauthorized interference with provider or other tenant assets. This action belongs to a different situation: A cloud customer requests a hypervisor disk image from a provider but the service contract grants access only to guest-level snapshots and audit events.
D:Here the decisive dependency is: Penetration testing in shared cloud environments must avoid unauthorized interference with provider or other tenant assets. This action belongs to a different situation: A cloud architect enables nested guest virtual machines without reviewing CPU virtualization extensions and hypervisor isolation requirements.
Question 392
Which specific remediation is the best controlled next step? A provider signs deployment artifacts but an organization uploads unverified public datasets into an automated fraud model every night. Select the ONE best answer.
- Perform privacy threat assessment, minimize training attributes and test model outputs for memorization or inference leakage before release.
- Select durable storage with documented retention, redundancy and restore support rather than instance-tied ephemeral scratch storage.
- Separate key administration from data access, enforce dual control for destructive operations and test key-recovery governance.
- Validate source provenance, integrity, schema and suspicious poisoning patterns before datasets enter training or inference pipelines.
Explanation
Correct Answer
D:Artifact authenticity alone does not prevent adversarial or low-quality input data from corrupting machine-learning behavior. Acceptance criterion: A poisoned test sample is quarantined with lineage evidence while vetted datasets continue producing reproducible outputs.
Incorrect Answers
A:Perform privacy threat assessment, minimize training attributes and test model outputs for memorization or inference leakage before release. Here the decisive dependency is: Artifact authenticity alone does not prevent adversarial or low-quality input data from corrupting machine-learning behavior. Use the specified intervention: Validate source provenance, integrity, schema and suspicious poisoning patterns before datasets enter training or inference pipelines.
B:That option would select durable storage with documented retention, redundancy and restore support rather than instance-tied ephemeral scratch storage, but it addresses a different concern. Artifact authenticity alone does not prevent adversarial or low-quality input data from corrupting machine-learning behavior.
C:Separate key administration from data access, enforce dual control for destructive operations and test key-recovery governance. Here the decisive dependency is: Artifact authenticity alone does not prevent adversarial or low-quality input data from corrupting machine-learning behavior. Use the specified intervention: Validate source provenance, integrity, schema and suspicious poisoning patterns before datasets enter training or inference pipelines.
Question 393
After the authorized change, which measurable result most directly demonstrates success? A team disables DNSSEC validation after noticing intermittent DNS failures but never checks whether signed records or resolver trust changed. Select the ONE best answer.
- A detected regression maps to one artifact digest and the approved rollback restores the previous security policy.
- The application signs authorized records without exporting private key bytes and key-usage events are attributable to scoped identities.
- A sample database restores within approved RTO and RPO boundaries and passes application consistency verification.
- Legitimate signed records resolve while tampered responses are rejected and resolver availability meets the requirement.
Explanation
Correct Answer
D:DNSSEC protects authenticity of DNS responses when its signatures and trust chain are correctly maintained and validated. It validates the controlled change: Diagnose validating resolver behavior, DNSSEC chains and time synchronization, then repair the trusted DNS configuration.
Incorrect Answers
A:The alternative observation is: A detected regression maps to one artifact digest and the approved rollback restores the previous security policy. The acceptance test for this case is: Legitimate signed records resolve while tampered responses are rejected and resolver availability meets the requirement.
B:An HSM protects key material and cryptographic operations within a defined trust boundary; application runtime secrets may expose raw keys. This is a different acceptance outcome; the required result here is: Legitimate signed records resolve while tampered responses are rejected and resolver availability meets the requirement.
C:Backup job success does not prove recoverability or application transaction consistency after a disruptive failure. This is a different acceptance outcome; the required result here is: Legitimate signed records resolve while tampered responses are rejected and resolver availability meets the requirement.
Question 394
Which action directly resolves the core infrastructure requirement without relying on an unrelated workaround? A cloud service incident is repeatedly closed once customer traffic recovers, even though the same certificate-expiry condition causes a weekly outage. Select the ONE best answer.
- Deploy scoped service identities, supported agent permissions, credential rotation and audit collection in the management installation workflow.
- Open a problem-management investigation to identify root cause and implement a lasting certificate lifecycle control beyond incident restoration.
- Normalize timestamps, document source clock offsets and collection completeness, and preserve original event records with verified integrity.
- Use controlled forensic acquisition and write-protected evidence storage with validated hashes and chain-of-custody transfers.
Explanation
Correct Answer
B:Before making it, verify: Compare incident timelines, certificate inventory, renewal automation, problem records and recurrence metrics.
Incorrect Answers
A:That option would deploy scoped service identities, supported agent permissions, credential rotation and audit collection in the management installation workflow, but it addresses a different concern. Incident management restores service, while problem management addresses recurring underlying causes and prevention.
C:Normalize timestamps, document source clock offsets and collection completeness, and preserve original event records with verified integrity. Here the decisive dependency is: Incident management restores service, while problem management addresses recurring underlying causes and prevention. Use the specified intervention: Open a problem-management investigation to identify root cause and implement a lasting certificate lifecycle control beyond incident restoration.
D:That option would use controlled forensic acquisition and write-protected evidence storage with validated hashes and chain-of-custody transfers, but it addresses a different concern. Incident management restores service, while problem management addresses recurring underlying causes and prevention.
Question 395
An architect must identify the primary underlying platform behavior. Which explanation is correct? A vendor submits an ISO/IEC 27017 certificate for a single regional datacenter but wants it treated as proof that all worldwide SaaS services meet the customer's controls. Select the ONE best answer.
- Independent attestations provide assurance only for systems and periods explicitly inside their stated boundary and criteria.
- Ethical and regulatory issues can persist even when the AI API is encrypted, available and technically accurate on average.
- Portability and reversibility require usable data and configuration outside a provider, not just nominal permission to download bytes.
- A platform control report does not automatically establish that a particular cryptographic implementation meets module validation requirements.
Explanation
Correct Answer
A:Evidence of correction: Supplier evidence maps to the purchased service scope and all uncovered customer requirements have documented additional evaluation.
Incorrect Answers
B:The alternative governing principle is: Ethical and regulatory issues can persist even when the AI API is encrypted, available and technically accurate on average. The mechanism tested by the reported case is: Independent attestations provide assurance only for systems and periods explicitly inside their stated boundary and criteria.
C:The alternative governing principle is: Portability and reversibility require usable data and configuration outside a provider, not just nominal permission to download bytes. The mechanism tested by the reported case is: Independent attestations provide assurance only for systems and periods explicitly inside their stated boundary and criteria.
D:The alternative governing principle is: A platform control report does not automatically establish that a particular cryptographic implementation meets module validation requirements. The mechanism tested by the reported case is: Independent attestations provide assurance only for systems and periods explicitly inside their stated boundary and criteria.
Question 396
Which action directly resolves the core infrastructure requirement without relying on an unrelated workaround? A managed generative-AI feature provides recommendations on employee eligibility without a process to review unfair outcomes or applicable disclosure duties. Select the ONE best answer.
- Validate training-data quality, current drift metrics, outcome feedback, and human review before retuning automated response thresholds.
- Specify an exit architecture with documented open export formats, data mapping, restore tests, and contractual assistance before migration.
- Institute an accountable AI risk process covering data minimization, human review, impact testing and jurisdiction-appropriate legal assessment.
- Adopt workload identities with scoped short-lived credentials, mutual authentication where supported, and least-privilege authorization.
Explanation
Correct Answer
C:Ethical and regulatory issues can persist even when the AI API is encrypted, available and technically accurate on average. Before making it, verify: Review model purpose, affected users, data categories, fairness test records, decision logs, and legal obligations.
Incorrect Answers
A:That option would validate training-data quality, current drift metrics, outcome feedback, and human review before retuning automated response thresholds, but it addresses a different concern. Ethical and regulatory issues can persist even when the AI API is encrypted, available and technically accurate on average.
B:Specify an exit architecture with documented open export formats, data mapping, restore tests, and contractual assistance before migration. Here the decisive dependency is: Ethical and regulatory issues can persist even when the AI API is encrypted, available and technically accurate on average. Use the specified intervention: Institute an accountable AI risk process covering data minimization, human review, impact testing and jurisdiction-appropriate legal assessment.
D:Adopt workload identities with scoped short-lived credentials, mutual authentication where supported, and least-privilege authorization. Here the decisive dependency is: Ethical and regulatory issues can persist even when the AI API is encrypted, available and technically accurate on average. Use the specified intervention: Institute an accountable AI risk process covering data minimization, human review, impact testing and jurisdiction-appropriate legal assessment.
Question 397
Before changing production, which focused diagnostic is most useful for identifying the true dependency? A global bank signs a master service agreement that does not state whether it can retrieve customer data and encryption keys at termination. Select the ONE best answer.
- Inspect SLA terms, maintenance exclusions, monitoring endpoints, remedies and example outage calculations.
- Review MSA, SOW, exit schedules, storage formats, key agreements, egress fees and a tested export plan.
- Inspect audit period, system boundaries, opinion, exceptions, migration timeline and updated control test records.
- Review quantified impact, risk appetite, acceptance authority, controls, decision records and scheduled risk review.
Explanation
Correct Answer
B:Cloud exit viability depends on practical rights and technical mechanisms to recover data and maintain control at contract termination. Negotiate clear data portability, deletion, key custody, assistance, timing and verification clauses before production lock-in. This check separates the suspected dependency before a risky change.
Incorrect Answers
A:An SLA percentage has limited value without measurable service boundaries and practical remedies for missed commitments. The focused check for the reported condition is: Review MSA, SOW, exit schedules, storage formats, key agreements, egress fees and a tested export plan.
C:Assurance reports address specified control periods and system descriptions, not future changes outside the covered scope. The focused check for the reported condition is: Review MSA, SOW, exit schedules, storage formats, key agreements, egress fees and a tested export plan.
D:Risk acceptance is a governance decision requiring informed authorized ownership rather than a technical team's unchecked preference. The focused check for the reported condition is: Review MSA, SOW, exit schedules, storage formats, key agreements, egress fees and a tested export plan.
Question 398
Which specific remediation is the best controlled next step? A confidential document is protected by certificate-based IRM, but an employee's certificate expires and authorized editing stops unexpectedly. Select the ONE best answer.
- Restrict token-vault access to approved workloads, isolate mapping data and log or approve sensitive detokenization operations.
- Select durable storage with documented retention, redundancy and restore support rather than instance-tied ephemeral scratch storage.
- Design certificate renewal and rights re-provisioning procedures with protected key continuity and emergency recovery options.
- Use supported information rights management with identity-scoped usage permissions, revocation and key expiry as applicable.
Explanation
Correct Answer
C:Acceptance criterion: Approved document access recovers after controlled certificate renewal without granting broad unreviewed privileges.
Incorrect Answers
A:Here the decisive dependency is: Rights enforcement may rely on certificate and key lifecycle, so validity and revocation directly affect authorized consumption. This action belongs to a different situation: A retailer replaces primary card numbers with tokens but puts the token-vault detokenization API behind an unrestricted service account.
B:Here the decisive dependency is: Rights enforcement may rely on certificate and key lifecycle, so validity and revocation directly affect authorized consumption. This action belongs to a different situation: A backup architect stores a regulated archive in ephemeral compute scratch disks because they appear cheaper than persistent object storage.
D:Here the decisive dependency is: Rights enforcement may rely on certificate and key lifecycle, so validity and revocation directly affect authorized consumption. This action belongs to a different situation: A consulting firm shares design documents externally and needs to revoke access to a former partner without destroying the master files.
Question 399
After the authorized change, which measurable result most directly demonstrates success? A cloud service incident is repeatedly closed once customer traffic recovers, even though the same certificate-expiry condition causes a weekly outage. Select the ONE best answer.
- The drift scanner detects a simulated insecure setting and restores the authorized state with a reviewed change record.
- A sample database restores within approved RTO and RPO boundaries and passes application consistency verification.
- A reviewer reproduces evidence digests and independently traces every copy and transfer without changing the source artifacts.
- The recurring outage stops after a validated certificate renewal mechanism and trend metrics confirm sustained resolution.
Explanation
Correct Answer
D:It validates the controlled change: Open a problem-management investigation to identify root cause and implement a lasting certificate lifecycle control beyond incident restoration.
Incorrect Answers
A:The alternative observation is: The drift scanner detects a simulated insecure setting and restores the authorized state with a reviewed change record. The acceptance test for this case is: The recurring outage stops after a validated certificate renewal mechanism and trend metrics confirm sustained resolution.
B:Backup job success does not prove recoverability or application transaction consistency after a disruptive failure. This is a different acceptance outcome; the required result here is: The recurring outage stops after a validated certificate renewal mechanism and trend metrics confirm sustained resolution.
C:Evidence handling must distinguish original acquired bytes from analysis copies and preserve verifiable acquisition history. This is a different acceptance outcome; the required result here is: The recurring outage stops after a validated certificate renewal mechanism and trend metrics confirm sustained resolution.
Question 400
Which specific remediation is the best controlled next step? A data science group deploys an AI model through a provider API but assumes provider infrastructure controls also validate customer prompts and sensitive training uploads. Select the ONE best answer.
- Define autoscaling rules using demand signals and a measured-service budget while documenting capacity and tenant isolation boundaries.
- Adopt workload identities with scoped short-lived credentials, mutual authentication where supported, and least-privilege authorization.
- Document the AI-as-a-service responsibility boundary and implement customer data validation, prompt handling, logging, and access governance.
- Institute an accountable AI risk process covering data minimization, human review, impact testing and jurisdiction-appropriate legal assessment.
Explanation
Correct Answer
C:Managed AI endpoints may secure model infrastructure while customers still control their submitted data, app behavior, and access policies. Acceptance criterion: Red-team test requests cannot expose other tenant data and uploaded datasets meet approved privacy and validation checks.
Incorrect Answers
A:That option would define autoscaling rules using demand signals and a measured-service budget while documenting capacity and tenant isolation boundaries, but it addresses a different concern. Managed AI endpoints may secure model infrastructure while customers still control their submitted data, app behavior, and access policies.
B:Adopt workload identities with scoped short-lived credentials, mutual authentication where supported, and least-privilege authorization. Here the decisive dependency is: Managed AI endpoints may secure model infrastructure while customers still control their submitted data, app behavior, and access policies. Use the specified intervention: Document the AI-as-a-service responsibility boundary and implement customer data validation, prompt handling, logging, and access governance.
D:Institute an accountable AI risk process covering data minimization, human review, impact testing and jurisdiction-appropriate legal assessment. Here the decisive dependency is: Managed AI endpoints may secure model infrastructure while customers still control their submitted data, app behavior, and access policies. Use the specified intervention: Document the AI-as-a-service responsibility boundary and implement customer data validation, prompt handling, logging, and access governance.