Beyond Eventual Consistency (FaunaDB, Calvin)

Have you ever wanted a distributed, highly-available database that provided ordered, transactional semantics?  With the release of Google Spanner to the public, Eric Brewer discussed the approach they took to achieve this goal, but it is only available via Google Cloud Platform.  Similar semantics can be found via FaunaDB, a database created to solve these very problems with the ideas of Daniel Abadi and others via the Calvin paper.  In this talk, we will discuss the theories behind Calvin, as well as the consistency/availability tradeoffs made in the face of partitions, as well as how we are planning to use it at Starbucks in our distributed service architecture.

Talk objectives:

• How modern distributed databases are tackling consistency without sacrificing availability.