Rapid Recall Key 34–65
- Merge applies survivorship and can change the effective master view; reversal may require reconstructing prior identities, values, mappings, and history.
- Global ID = enterprise reconciled identifier; Source ID = local application identifier.
- X-Ref preserves mapping/history between Source IDs and Global ID, including merge/unmerge changes.
- One authorized generator protects enterprise uniqueness and prevents competing/colliding Global IDs.
- Affiliation = flexible typed association among entities; parent-child = direct hierarchical relationship.
- Sharing/Stewardship adds human accountability, exception handling, validation, feedback, maintenance, publication, and continuous improvement.
- Registry: enterprise identity/index points to Master Data still maintained in source SORs; minimal source disruption but distributed assembly can be complex.
- Transaction Hub: central hub owns updates and becomes SOR; applications maintain/access Master Data through it.
- Consolidated: source systems remain SORs while copies are reconciled into a central System of Reference; replication/latency are trade-offs.
- Registry points to distributed source-held data; Consolidated stores a shared integrated copy.
- Transaction Hub centralizes writes/SOR; Consolidated leaves writes in source SORs and centralizes a replicated consumer view.
- Six MDM activities: drivers/requirements; source assessment; architecture; Master model; stewardship/maintenance/DQ/matching controls; governance + share/maintain/monitor. Equivalent chapter-consistent wording earns credit.
- Six RDM activities: requirements/authoritative sets; source assessment; RDM architecture/acquisition; model code sets/mappings/hierarchies/Metadata; stewardship + controlled change/history; governance + publish/share/monitor.
- RDM architecture emphasizes update frequency/volatility, external/vendor delivery, history/versions, consumption, manual stewardship, approvals/notifications, mappings—and does not need entity-resolution machinery like MDM.
- Steward, source authority, update schedule/frequency, consumers, version/effective dates, history needs, definitions, mappings, status/deprecation.
- A central RDM repository creates value only when policy, integrations, stewardship, and consuming systems actually use its governed values.
- Specialized MDM applications, database/integration/DQ technology, workflow/stewardship tools, sharing hubs, and combinations of enterprise platforms; technology supports the discipline.
- Incremental implementation manages complexity, prioritizes business value, enables governance/process learning, and delivers under an enterprise roadmap rather than a big bang.
- Organization structure; number/distribution of SORs; governance maturity; latency/access needs; consumers; integration complexity; volumes/DQ; desired write authority; cost/source-change feasibility.
- Movement monitoring shows lineage, actual sharing/use, root-cause paths, ingestion/latency, and integration/transformation effectiveness.
- Request → stakeholders → impact → decide/approve → update → communicate/inform.
- Examples: add value; change value/description; deprecate/remove value; change mappings/cross-references; change hierarchy/structure/granularity or adopt new external version.
- Planned = expected/scheduled such as annual standard update; ad hoc = unplanned and needs case-specific impact/approval.
- Shared data, permitted uses, provider/consumer duties, privacy/security/retention, DQ/service expectations, issue/escalation, accountability for changes/failures.
- SLAs make availability/timeliness/service expectations explicit and measurable between provider and consumer.
- Departments can resist surrendering local definitions, identifiers, values, and process control to enterprise governance.
- Governance decides authoritative sources, stewardship, DQ/conditions of use, approval gates, sharing, privacy/security/retention, monitoring, exceptions, and escalation.
- DQ/compliance; change activity; ingestion/consumption; SLA/service; steward coverage; TCO; sharing volume/usage/adoption.
- Steward coverage identifies governed domains/data sets with or without accountable stewards and exposes ownership gaps.
- Change activity shows frequency/type/source/lineage of changes and can reveal instability, workload, risk, or tuning needs.
- Ingestion/consumption shows providers, consumers, volume/use, and whether the shared capability is actually adopted.
- Final model: govern classifications through RDM; reconcile entity identity through MDM; define authority/architecture; preserve quality, identifiers, mappings, history, reversibility; distribute via stewardship/sharing; measure and improve.
Source: Chapter 10, pp. 329–359.