backward compatibility
The property that a new version of an interface still works for clients written against the old one.
In more detail
Additive changes, such as a new optional field or option, are usually backward compatible, but only if clients ignore what they do not recognise and old meanings stay the same. Under semantic versioning, compatible changes can ship in minor versions, while anything that could break a consumer needs a major version.
Example
Schema registry BACKWARD level: consumers on the new schema can read old records, so upgrade consumers first.Related terms
- breaking change — A change that makes existing users of a program or interface stop working until they change something too.
- additive change — Evolving an API only by adding things, such as a new optional field or operation, while everything existing keeps its meaning.
- API versioning — Giving an API contract a version so that incompatible changes can be released without breaking clients that use the old one.
- compatibility level — A schema registry setting per subject that decides which schema changes are allowed: BACKWARD, FORWARD, FULL or NONE.
- forward compatibility — A schema change is forward compatible when consumers still using the old schema can read data written with the new one.
- schema evolution — Changing the structure of your data over time while keeping existing producers and consumers working.
- semantic versioning — A version numbering scheme of three numbers, major.minor.patch, raised for breaking changes, new features and fixes respectively.
- rolling update — Replacing instances of an application with the new version a few at a time while the rest keep serving.
- database migration — A scripted, versioned change to a database's structure, such as adding a table or a constraint.
- schema — A formal description of the structure of data: which fields or columns exist, their types and which are required.
Taught in
- Naming Resources And Collections — Modern API Architecture: REST, GraphQL & gRPC
- Versioning APIs — Modern API Architecture: REST, GraphQL & gRPC
- Contract Testing — Modern API Architecture: REST, GraphQL & gRPC
- The Field The Registry Refused — Confluent Apache Kafka Certifications (CCDAK and CCAAK) Exam Prep
- Seventy One Percent — Confluent Apache Kafka Certifications (CCDAK and CCAAK) Exam Prep
- System Governance: Changesets, Semantic Versioning & Deprecation — Design Systems & Component Engineering
- Deployment Strategies: Rolling Updates — DevOps & Site Reliability Engineering
- Deployment Strategies: Blue Green — DevOps & Site Reliability Engineering
- Compatibility Modes — Apache Kafka & Event Streaming
- Evolving An Event Without Breaking Consumers — Apache Kafka & Event Streaming
References
- Google Cloud API design guide: Compatibility
- Confluent: schema evolution and compatibility
- Semantic Versioning 2.0.0
- Danilo Sato on martinfowler.com: ParallelChange