PCCN-K (Knowledge Professional) flashcards

32 cards. Read the term and the definition here. The deck at the top of this page is optional.

NIST essential characteristics

Five traits defining cloud: on-demand self-service, broad network access, resource pooling, rapid elasticity, and measured service. A model missing measured service and elasticity is not cloud in the NIST sense.

Shared responsibility model

Division of security duties between provider and customer. The provider secures facilities, hardware, and hypervisor; the customer secures identity, data, and configuration. The split shifts toward the customer from SaaS to IaaS.

Service catalog purpose

Authoritative list of services offered, with tiers, owners, prices, and SLAs. It governs what users may request; the CMDB separately records what is actually deployed and how items relate.

Cloud service lifecycle stages

Strategy, design, transition, operation, and continual improvement. Each stage produces defined outputs and has an accountable role, so handoffs between stages are auditable rather than informal.

Multi-tenancy risk

Risk from shared infrastructure: noisy-neighbor contention, residual data on reused storage, and side-channel exposure. Mitigation requires both technical isolation controls and contractual guarantees from the provider.

Utility vs capital expenditure

Cloud converts upfront capital spend into usage-based operating spend. This requires metering, rating, and allocation capability so costs can be attributed to the consuming business unit.

Service model control shift

As you move from IaaS to PaaS to SaaS, the provider manages more of the stack and the customer manages less. Responsibility for identity, data, and access configuration typically stays with the customer.

Six Rs migration strategies

Retire, retain, rehost, replatform, refactor, and repurchase. Each option carries different cost, risk, and timeline, so the strategy output is a portfolio decision rather than a single approach.

Governance decision rights

Explicit assignment of who approves new cloud services, owns policy, and accepts residual risk. Unclear ownership is a governance defect, not an operational issue to be resolved later.

Policy traceability chain

Regulation flows to enterprise policy, then to a cloud control baseline, then to service configuration. Each lower layer must be traceable to a higher requirement or it cannot be justified during audit.

FinOps three phases

Inform, optimize, operate. Inform gives visibility and cost allocation; optimize covers rightsizing and commitment coverage; operate runs budgets, anomaly alerts, and showback or chargeback to consumers.

Exit plan requirements

Documented data export format, maximum extraction time, and cost of leaving. Addresses lock-in from proprietary APIs, egress charges, identity integration, and licensing before the contract is signed.

Budget variance formula

Variance percent equals actual minus budget, divided by budget, times one hundred. A positive result is overspend. Report at tag, account, and service level for meaningful attribution.

Commitment coverage ratio

Committed spend divided by total eligible spend. Target coverage depends on workload stability; over-committing on volatile workloads creates waste rather than savings.

Availability of serial components

For components in series, multiply availabilities: two at 99% yield 98.01%. Adding dependencies always lowers end-to-end availability, so unnecessary serial hops should be removed.

SLA versus SLI versus SLO

SLA is the contractual commitment with remedies; SLI is the measured indicator; SLO is an internal target set tighter than the SLA so breaches are detected before customers are affected.

Error budget calculation

Error budget equals one minus the SLO. A 99.9% monthly SLO permits roughly 43 minutes 50 seconds of unavailability in a 30-day month, which teams may spend on releases.

Release versus deployment

Deployment moves code into an environment; release exposes it to users. Decoupling them enables canary, blue/green, and feature-flag patterns that limit blast radius.

Infrastructure as code drift

Drift is divergence between declared and deployed state. Detection compares the two and must trigger remediation or a recorded exception, otherwise the code is no longer the source of truth.

Transition acceptance criteria

Measurable conditions that must be met and signed before operations accepts a service. Vague criteria shift unresolved defects onto the operations team after go-live.

Rollback trigger definition

Pre-agreed thresholds, such as error rate or latency, plus maximum tolerable rollback time. Irreversible data changes require a forward-fix plan instead of rollback.

Incident priority derivation

Priority comes from impact multiplied by urgency, not from the seniority of the reporter. Severity level then determines response time and communication cadence.

Major incident process

Declare the incident, appoint a commander, set a communication cadence, restore service first, and perform root-cause analysis only after stabilization. Investigation must not delay restoration.

Standard change definition

A pre-authorized, low-risk change with a documented procedure, so it needs no per-instance approval. Emergency changes are expedited and reviewed retrospectively.

Incident versus problem

Incident management restores service; problem management removes the underlying cause. Recurring incidents with no problem record indicate the process is failing.

RPO versus RTO

Recovery point objective is acceptable data loss measured in time; recovery time objective is acceptable downtime. Backup restores data, disaster recovery restores service.

Availability percentage formula

Agreed total minutes minus unplanned downtime, divided by agreed total minutes, times one hundred. Maintenance windows count as downtime unless the contract explicitly excludes them.

MTTR and MTBF relation

MTTR is total downtime divided by incident count; MTBF is total uptime divided by failure count. Availability approximates MTBF divided by MTBF plus MTTR.

Percentile latency reporting

The p95 latency is the value below which 95% of requests fall. Percentiles expose tail latency that averages hide, so they are preferred for SLA and SLO reporting.

SLA credit mechanics

Credits are tiered remedies for shortfall, usually requiring a customer claim within a defined window. They compensate partially and do not restore the lost service.

Cost per transaction metric

Total service cost divided by transaction volume. Used to compare service tiers and to detect unit-cost drift as the estate scales.

Risk treatment options

Avoid, reduce, transfer, or accept. Acceptance must be time-bound, recorded in the risk register, and owned by an accountable person who reviews it at expiry.