Notes on Consistency Models

1 minute read

NOTE These post series is merely a backup of some notes from my PhD sometime in 2016

In Consistency models, the safety property restricts the set of possible values observed by clients accessing different replicas.

The set of possible executions is called a consistency model. A consistency model is said to be stronger than another model if its executions are a proper subset of the executions of the other.

Causal Consistency ensures that the causes of an operation are visible at a replica no later than the operation itself.

Eventual Consistency is a liveness property, which is described by update propagation[2]. Eventual consistency does not state how the system becomes consistent – only that replicas are allowed to diverge.

Strong Eventual Consistency defines the limits of divergence of Eventual Consistency. It specifies that replicas that receive the same set of updates will converge to the same state. This is the consistency model used/offered in the CRDT framework (Conflict-Free Replicated Data Types).


  • Hagit Attiya, Faith Ellen, and Adam Morrison. 2015. Limitations of Highly-Available Eventually-Consistent Data Stores. In Proceedings of the 2015 ACM Symposium on Principles of Distributed Computing (PODC ’15). ACM, New York, NY, USA, 385-394. DOI=
  • Sebastian Burckhardt, Daan Leijen, Jonathan Protzenko, and Manuel Fähndrich. Global Sequence Protocol: A Robust Abstraction for Replicated Shared State. In ECOOP ’15.

Leave a Comment

Your email address will not be published. Required fields are marked *