Materialize is a streaming database. It ingests data from sources (i.e., Kafka, Postgres CDC), incrementally maintains materialize views, and allows users to query or subscribe the data using Postgres wire protocol or sinks data to sinks like Kafka.
- Website
- https://materialize.com/[01]
- Tech Docs
- https://materialize.com/docs/[03]
- @MaterializeInc
- Developer
- Country of Origin
- US
- Start Year
- 2019 [06]
- Project Type
- Commercial
- Written in
- Rust
- Compatible With
- PostgreSQL
- Operating System
- Linux
- License
- Business Source License
Materialize is a streaming database. It ingests data from sources (i.e., Kafka, Postgres CDC), incrementally maintains materialize views, and allows users to query or subscribe the data using Postgres wire protocol or sinks data to sinks like Kafka.
History[04]
Materialize builds on top of differential dataflow, which is a differential computation framework for timely dataflow, introduced in Naiad system from Microsoft Research.
Before the v0.27 release, Materialize mainly supports in-memory computation. Materialized views are only maintained in memory. It requires the data source to persist and replay the changes after restarting Materialize. On October 12, 2022, Materialize 0.26 LTS entered long-term-support mode and 0.27 debuted as the cloud-native release of Materialize. It added support for source data persistence, materialized views with data replicated and persisted within cluster, and other changes for the cloud platform.
Materialize is licensed under BSL license. Each version of Materialize will be transformed to Apache 2.0 License 4 years after its release.
Indexes[05]
Differential dataflow stores changes of a sequence of data (collection) by key and time in an in-memory index called arrangement, so that operators can query the changes of value for a given key in a period of time.
The index is maintained in-memory and provides multi-version access to the data.
Citations
6 sources- The Live Data Layer for Apps and AI Agents | Materialize materialize.com
- GitHub - MaterializeInc/materialize: The live data layer for apps and AI agents. Create up-to-the-second views into your business, just using SQL · GitHub github.com
- Home | Materialize Documentation materialize.com
- About Us & Our Values | Materialize materialize.com
- https://etos.cs.brown.edu/publications/papers/shared-arrangements:vldb20.pdf brown.edu
- Initial commit github.com