Certified Technical Consultant flashcards

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

Scope baseline

The signed statement of work plus all approved change requests. Any requested deliverable not traceable to this baseline is out of scope until a formal change order is approved.

Cost Performance Index

CPI equals earned value divided by actual cost. A value below 1.0 means the engagement is spending more than the work completed is worth, signaling a cost overrun.

Schedule Performance Index

SPI equals earned value divided by planned value. A value below 1.0 indicates the project is behind schedule relative to the baseline plan at the measurement date.

RACI matrix

A responsibility assignment grid. Each activity has exactly one Accountable owner, one or more Responsible performers, Consulted contributors engaged before decisions, and Informed recipients notified afterward.

Change control sequence

Log the request, assess impact on scope, schedule, cost and risk, obtain written approval from the authorized client sponsor, rebaseline the plan, then execute the approved change.

Non-functional requirement

A quality attribute describing how well a system must perform, such as latency, availability, throughput, scalability or security. It must be measurable to be testable and verifiable.

Fit-gap analysis

Classification of each requirement as standard fit, configuration fit, customization, or process change. The count of customizations is the main driver of total cost of ownership.

Requirements traceability matrix

A table linking each requirement identifier to its business justification, design element, build component, and test case. Orphan requirements lack either upstream justification or downstream verification.

MoSCoW prioritization

Ranks requirements as Must, Should, Could, or Won't have. Must items must together form a viable deliverable solution, and Won't items are explicitly deferred from the current release.

Testable requirement criterion

A requirement is testable only when it states a measurable condition and expected outcome. Vague terms such as fast, intuitive or user-friendly are defects requiring clarification.

Idempotent message consumer

A consumer designed so repeated delivery of the same message produces the same end state. Achieved with natural keys, deduplication stores, or upsert operations rather than blind inserts.

Synchronous versus asynchronous integration

Synchronous request and response blocks the caller and fails immediately. Asynchronous messaging decouples producer and consumer, tolerating outages but requiring retries, ordering rules and dead-letter queues.

Authentication versus authorization

Authentication proves who a user is, typically through SAML or OIDC. Authorization determines what that authenticated user may do, using roles, scopes or policy rules. They are separate controls.

CAP theorem

Under a network partition, a distributed system must sacrifice either consistency or availability. The chosen trade-off is an explicit architecture decision with business consequences.

Recovery point and time objectives

RPO is the maximum acceptable data loss measured in time; RTO is the maximum acceptable downtime. Both must be quantified and validated by an actual failover test.

Configuration versus customization

Configuration uses supported platform settings and survives upgrades. Customization requires code or schema change and carries upgrade, testing and maintenance cost. Prefer configuration where requirements allow.

Data migration reconciliation

After loading, compare source and target record counts, key field totals and checksums. Unreconciled differences block sign-off and must be resolved or formally accepted before cutover.

Environment promotion path

The sequence development, integration test, staging or UAT, then production. Each promotion uses a versioned, repeatable deployment artifact rather than manual ad hoc changes.

Rollback plan

A pre-approved procedure to restore the prior state if deployment fails, using reverse scripts or a verified restore point tested in non-production before the production change window.

Bulk load rate limits

Platform APIs cap requests per interval. Bulk loads must batch, throttle and monitor errors, because parallel writes to the same records cause lock contention and partial failures.

Smoke versus regression testing

Smoke testing quickly confirms a new build is stable enough for deeper testing. Regression testing re-executes existing cases after any change to detect broken previously working behavior.

Defect severity versus priority

Severity measures impact on the system or business; priority measures urgency of fixing. They are independent attributes and frequently disagree, so both fields are tracked separately.

UAT exit criteria

Predefined conditions ending user acceptance testing: all critical and high defects closed, agreed defect tolerance met, test cases executed, and business sign-off formally recorded.

Parallel run deployment

Both legacy and new systems operate simultaneously for a defined period, with outputs compared. It is the safest and most expensive cutover strategy and reduces transition risk.

Go/no-go readiness checklist

The gate before production cutover: test completion, data reconciliation, training completion, confirmed rollback plan, support rota, and named business approver recorded in writing.

Hypercare period

A defined post-deployment window of intensified monitoring and rapid issue resolution, tracking error rates, integration queue depth and user reports before transition to steady-state support.

RAG status criteria

Red, amber, green status must map to stated thresholds, such as red meaning a milestone slip over five working days or an unresolved critical issue. Undefined colors mislead sponsors.

Decision log

A record of each significant decision with date, decision, rationale, alternatives rejected and owner. It resolves later disputes about why the solution was designed as it is.

Handover package contents

As-built documentation, configuration workbook, operational runbook, support model, transferred credentials and warranty terms. It enables the client to operate and maintain the solution independently.

Knowledge transfer verification

Confirming the client can perform the task unaided, observed in a practical exercise. Attendance at a training session alone does not demonstrate transferred competence.

Data processing agreement

A contract required when a processor handles personal data for a controller. It sets purpose limits, security measures, sub-processor rules, breach notification and deletion obligations.

Conflict of interest disclosure

Any personal, financial or competing interest that could bias advice must be disclosed in writing as soon as identified, allowing the client to decide on mitigation or continuation.