CCSO flashcards
CCSO · Build recall with a focused review session.
32 cards. Read the term and the definition here. The deck at the top of this page is optional.
- Shared responsibility model
- Division of security duties between cloud provider and customer. The provider secures physical infrastructure and virtualization; the customer secures data, identity, configuration, and workload-level controls regardless of service model.
- Control inheritance
- Reuse of provider audit attestations, such as SOC 2 or ISO 27001 reports, to satisfy customer framework requirements. Inherited controls still need customer-side configuration evidence to be audit-valid.
- Residual risk
- Risk remaining after controls are applied. It is compared against the organization's stated risk appetite, and acceptance above appetite requires documented management sign-off.
- Policy as code
- Expressing security requirements as machine-evaluated rules enforced in pipelines and at deploy time. Preventive policies deny noncompliant resources; detective policies flag drift after provisioning.
- Data sovereignty
- Requirement that data remains subject to the laws and access rules of the jurisdiction where it is stored. It constrains region choice, replication targets, and support personnel access.
- Trust boundary
- Any point where the level of control or authentication changes, such as account-to-account, network-to-internet, or control-plane-to-data-plane. Every boundary needs explicit authentication and authorization.
- Blast radius
- Scope of impact if a component is compromised. Account isolation, network segmentation, and least privilege bound it; flat single-account architectures maximize it.
- Control plane vs data plane
- Management APIs versus workload traffic. Control-plane compromise usually yields full resource takeover, so it warrants stronger authentication, logging, and network restriction than data-plane access.
- Landing zone
- Pre-built, governed multi-account foundation providing identity, network segmentation, centralized logging, and guardrails. Workloads deploy into it so compliance is structural rather than retrofitted.
- Federated identity
- Authentication delegated to an external identity provider via SAML or OIDC. The cloud provider stores no passwords, and disabling the identity at the IdP revokes cloud access immediately.
- STS temporary credentials
- Short-lived tokens issued in exchange for an authenticated identity or role assumption. They expire automatically, reducing the exposure window compared with long-lived access keys.
- Permissions boundary
- A maximum-permissions ceiling attached to an identity. Effective permissions are the intersection of identity policies and the boundary, preventing privilege escalation even if a policy is overly broad.
- Just-in-time privilege
- Elevated access granted on request, with approval and time-bound expiry, instead of standing admin rights. Sessions are logged and recorded to support audit and separation of duties.
- Break-glass account
- Sealed emergency credential used when normal federation fails. It is monitored, alerts on use, tested periodically, and excluded from routine access paths.
- Access recertification
- Periodic owner attestation that each entitlement is still required. Unused roles and orphaned accounts are removed during review rather than deferred to a later cycle.
- Envelope encryption
- A data key encrypts the data, and a root key encrypts the data key. Rotating the root key does not require re-encrypting stored data, only rewrapping data keys.
- Customer-managed key
- Encryption key the customer creates and controls in the cloud key service. It enables independent audit of key use and revocation, which can render protected data unreadable.
- Crypto-shredding
- Destroying encryption keys so ciphertext becomes permanently unrecoverable. It satisfies deletion obligations only when key destruction is provable and all backups using that key are covered.
- Tokenization
- Replacing sensitive values, such as card numbers, with non-sensitive surrogate tokens that preserve referential integrity. It reduces compliance scope because the original data leaves the environment.
- Stateful vs stateless firewall
- Security groups are stateful and remember connections, so return traffic is allowed automatically. Network ACLs are stateless and evaluate each direction separately using numbered allow and deny rules.
- Private endpoint
- An interface or gateway that routes service traffic over the provider's private network instead of the public internet. It reduces exfiltration exposure and removes public attack surface.
- Micro-segmentation
- Identity-aware policy applied between individual workloads rather than at the network perimeter. It limits east-west lateral movement after an initial host compromise.
- Default-deny egress
- Outbound traffic blocked unless explicitly allowed by destination and port. It disrupts data exfiltration and command-and-control callbacks that rely on unrestricted outbound connections.
- Admission control
- A cluster gate that evaluates workloads before scheduling and rejects those violating policy, such as unsigned images or missing resource limits. It is preventive rather than detective.
- Immutable image digest
- A content hash identifying an exact container image version. Deploying by digest instead of a mutable tag prevents tag-mutation attacks and supports reproducible, verifiable provenance.
- SBOM
- A software bill of materials listing components and versions in an artifact. It enables rapid identification of affected workloads when a dependency vulnerability is disclosed.
- Log immutability
- Write-once retention, often via object lock, applied to centralized audit logs. It prevents a compromised workload or insider from deleting evidence of their own activity.
- MTTD and MTTR
- Mean time to detect and mean time to respond. They are the core operational security metrics, and gaps in log coverage directly increase detection time.
- Incident containment
- Limiting incident spread through credential revocation, key rotation, or network isolation. Evidence such as snapshots and logs must be preserved before remediation destroys it.
- Alert fatigue
- Analyst desensitization caused by high volumes of low-value alerts. It is measured by the actionable-alert ratio and mitigated by tuning thresholds and suppressing known-benign patterns.
- Shift left
- Moving security checks into design, code review, and build so defects are caught before deployment. It complements, rather than replaces, runtime detection and response.
- Drift detection
- Continuous comparison of running configuration against the declared infrastructure-as-code baseline. Detected differences trigger alerts or automated remediation and are logged with their trigger.