README.md 1.3 KB

Materialize Platform

Materialize Platform is the evolution of materialized into a horizontally scalable, highly available distributed system.

The intended experience with Materialize Platform is that many independent users can establish changing data sets, create and interrogate views over these data, and share the data, views, and results with each other in a consistent fashion. The intent is that in each dimension we remove limits to scaling:

  • We aim to support unbounded numbers of concurrent users and sessions
  • We aim to support unbounded numbers of data sources, and with unbounded rates
  • We aim to support unbounded numbers of views over these data

Here "unbounded" does not mean "infinite", only "can be increased by spending more resources".

It is important to stress that while we want to provide access to more resources, and more availability, this is not meant to come at the expense of correctness or consistency. We still require Materialize to speak clearly and correctly about the contents of the collections it manages.

The above are goals, and the path to Materialize Platform starts with none being the case yet.

Initial design work has begun in the documents below.