Glossary Updates12 new terms added to the glossaries · October 2, 2026, 22:44 CEST
AI TechDocKnowledge

Glossary · System coordination, integration and orchestration

Change data capture (CDC)

Also known as: CDC, change-data capture

German: Änderungsdatenerfassung (CDC)

In data integration, change data capture (CDC) is a technique that detects inserts, updates and deletes in a source database, typically by reading its transaction log, and delivers them as a stream of change events to other systems.

  • System integration

In one sentence

Change data capture detects inserts, updates and deletes in a source database and streams them as change events to other systems.

Example

Every change to the material master in the ERP database is captured by CDC and published to the event broker, so the MES updates its copy within seconds.

How it applies

  • Integration: CDC avoids repeated bulk exports and puts little extra load on the source application. Consumers receive changes in commit order, which supports Data replication and event-driven integration.
  • Engineering: Plan for schema changes in the source, for consumers that fall behind and for the initial load before change streaming starts. Consumers of CDC data are usually eventually consistent (see Eventual consistency) with the source.
  • Documentation: Document which tables are captured, which system is the Source of truth for each data item and how long change history is retained. Personal data in captured tables may need separate protection.

CDC vs. data replication

CDC is a way of detecting and delivering changes. Data replication is the broader goal of keeping copies of data in sync; it can use CDC, snapshots or application-level messages.

By knowledge.aitechdoc.world · Published September 26, 2026 · Last reviewed

Source: AI TechDoc Blog editorial definition, based on data integration practice

Definitions follow the cited standards and specifications. Where a source is a copyrighted publication, such as an ISO, IEC or EN standard, the definition is a close paraphrase, not a verbatim quotation, so as not to infringe copyright. We recommend reading the original publication. The sections “How it applies” are editorial commentary by AI TechDoc Blog and are not part of any standard.

Seen a mistake? Send us a note!