A Short Primer on Causal Consistency

The growing prevalence of geo-distributed services that span multiple geographically separate locations has triggered a resurgence of research on consistency for distributed storage. The CAP theorem and other earlier results prove that no distributed storage system can simultaneously provide all desirable properties—e.g., CAP shows this for strong Consistency, Availability, and Partition tolerance—and some must be sacrificed to enable others. In this article, we suggest causal consistency represents an excellent point in this tradeoff space; it is compatible with strong performance and liveness properties while being far easier to reason about than the previously-settled-for choice: “eventual” consistency.